survey of open source integrated library systems - sjsu scholarworks
Post on 03-Feb-2022
1 Views
Preview:
TRANSCRIPT
San Jose State UniversitySJSU ScholarWorks
Master's Theses Master's Theses and Graduate Research
2008
Survey of open source integrated library systemsLinda M. RieweSan Jose State University
Follow this and additional works at: http://scholarworks.sjsu.edu/etd_theses
Part of the Library and Information Science Commons
This Thesis is brought to you for free and open access by the Master's Theses and Graduate Research at SJSU ScholarWorks. It has been accepted forinclusion in Master's Theses by an authorized administrator of SJSU ScholarWorks. For more information, please contact scholarworks@sjsu.edu.
Recommended CitationRiewe, Linda M., "Survey of open source integrated library systems" (2008). Master's Theses. 3481.http://scholarworks.sjsu.edu/etd_theses/3481
SURVEY OF OPEN SOURCE INTEGRATED LIBRARY SYSTEMS
A Thesis
Presented to
The Faculty of the School of Library and Information Science
San Jose State University
In Partial Fulfillment
of the Requirements for the Degree
Master of Library and Information Science
by
Linda M. Riewe
August 2008
UMI Number: 1459712
INFORMATION TO USERS
The quality of this reproduction is dependent upon the quality of the copy
submitted. Broken or indistinct print, colored or poor quality illustrations and
photographs, print bleed-through, substandard margins, and improper
alignment can adversely affect reproduction.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if unauthorized
copyright material had to be removed, a note will indicate the deletion.
®
UMI UMI Microform 1459712
Copyright 2008 by ProQuest LLC.
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.
ProQuest LLC 789 E. Eisenhower Parkway
PO Box 1346 Ann Arbor, Ml 48106-1346
APPROVED FOR THE SCHOOL OF LIBRARY AND INFORMATION SCIENCE
Dr. Daniel W. Fuller
•[is-- y~£ Dr. William Fisher
{ — • — - •
Steve Perry, M.A
APPROVED FOR THE UNIVERSITY
j%* / M^^r^>^ (fifm?
ABSTRACT
SURVEY OF OPEN SOURCE INTEGRATED LIBRARY SYSTEMS
by Linda M. Riewe
The purpose of this study is to compare integrated library system (ILS) costs and
benefits, and to inform librarians about considerations when choosing between an open
and a closed source ILS. A survey was conducted among libraries that used the largest
open source ILSs, Koha and Evergreen, and various proprietary ILSs. Questions were
asked about initial and annual costs. Questions about benefits took the form of
satisfaction ratings. The questionnaire also asked why the ILS was selected, whether it
was customized, who hosted and serviced it, what problems were encountered, and
library demographics.
The survey found that open source ILSs were more cost-effective than
proprietary ILSs. Libraries using open source ILSs chose them mainly for affordability,
and they cost less than proprietary ILSs. Although users of open source ILSs
experienced difficulties with installation and incomplete documentation, they were
modestly more satisfied than users of proprietary ILSs.
ACKNOWLEDGMENTS
I would like to thank my parents, who remain with me in spirit.
Many thanks are due to my thesis committee members, Dr. Daniel Fuller,
Dr. William Fisher, and Steve Perry, who read the manuscript during the summer while
they were not paid. Steve Perry also supervised three units of this research for which he
was not paid.
1 would like to express my gratitude to Silvio Levy and Sheila Newbery for their
support, to David Robbins for his help with editing, and to Elizabeth Garcia for her
interview about Evergreen.
Last but not least, thanks to Chuck, Bongsook (-o"^r), Eugene, and Hannah, for
being themselves.
v
Table of Contents
Introduction 1 General Description of Area of Interest 1 Problem to be Investigated 5 Significance of Problem and Justification 6
Literature Review 8 History 8 Theory Identified as Framework for Research 16 Open Information Models 17
Traditional principles and open information models 18 Patents are incompatible with open source 20 Open standards 22 Open access 23 Mandates for public access to publicly funded research 27 Summary 28
Current Research Findings and Gaps 29 Koha 31 Evergreen 33
Research Procedures 35 Survey Introduction 35 Sample and Recruitment 36 Survey Method 37 Respondent Characteristics 39
Findings 43 Reasons for ILS Choice 43 ILS Satisfaction Ratings 44 ILS Problems 48 ILS Costs 50
Evergreen costs 50 ILS costs estimated by vendors 52 Comparison of open source and proprietary ILS costs 53
Summary of the Findings 56 Conclusion 58
Alternative Explanation and Limitations of the Study 58 Recommendations for Future Research 59 Implications for Future Practice 61 Discussion 62 Summary 65
References 67 Appendix A: Survey Questionnaire 77 Appendix B: Tables B1-B6 87 Appendix C: Glossary of Software Classified by Category of Intellectual Property 93
VI
List of Tables
Table 1. Software Benefits 5 Table 2. Library Size 39 Table 3. Library Type 40 Table 4. Library Location 41 Table 5. ILS Hosting and Service Providers 42 Table 6. Top Five Reasons for ILS Choice 43 Table 7. Annual Cost of Previous ILS at Georgia PINES Consortium 50 Table 8. Evergreen Size and Costs at Georgia PINES Consortium 51
Appendix B
Table Bl. ILS Satisfaction Ratings 87 Table B2. System Prerequisites Cost of ILS 88 Table B3. Initial Cost of ILS 89 Table B4. Annual Cost of ILS 90 Table B5. Initial Materials Cost of ILS 91 Table B6. Initial Labor Cost of ILS 92
vn
Introduction
General Description of Area of Interest
The main type of software in use in libraries today is the integrated library
system (ILS), which is the modern equivalent of the card catalog. An ILS provides a
search interface to the library catalog and automates library tasks such as the tracking of
book loans and returns. Although ILS vendors have added many different features,
every ILS has nearly the same core components of cataloging and circulation tracking.
Because the ILS core is stable, it is suitable for collaborative development.
Collaboratively developed computer programs are known as free software or open
source software (OSS). Librarians and programmers have worked together to produce
several open source ILSs. Users and developers are free to share and change open
source programs, a practice similar to sharing recipes. In OSS the sequence of humanly
readable computer instructions, known as source code, is open to view. Open source
licenses ensure that OSS and its derivatives may be freely viewed, used, copied,
modified, and redistributed (Open Source Initiative, 2006). Examples of well-known
OSS include the Mozilla Web browser, Apache Web server, and Linux operating system.
Open source ILSs are new compared with proprietary ILSs and share less than
two percent of the ILS market (Breeding, 2007a; Pace, 2005). Vendors of proprietary
ILSs do not allow users to view the source code, copy, or customize it. However,
proprietary ILSs are mature turnkey products whose use requires little technical
knowledge. OSS is of potential interest to librarians because libraries might save money
by using it, they may customize it, and because the open source community and
1
librarians share the same values of equal access and freedom of information.
The purpose of the study is to inform librarians about what considerations to
make when choosing between an open and a closed source ILS. The study investigates
why some libraries have chosen an open source ILS and compares the costs and benefits
of the open source ILS with those of a proprietary ILS. Librarians using both open
source and closed source ILS software will be surveyed.
The precursor of OSS was free software, meaning/ree in the sense of liberty to
view and modify the source code. Users were free to participate in developing and
changing the software. Advocates of free software emphasized its freedoms and
unfettered access. In English, the word free also means no cost, but this is not its
intended meaning in the term free software. A group of software developers and users
later adopted the term open source because it is more business-friendly (Williams, 2002,
chap. 11). Despite their origins, both terms have come to connote low cost. Free
software and open source software are nearly the same for practical purposes. Both are
often licensed under the same General Public License (Free Software Foundation, 1991).
A term containing both concepts is free/open source software (F/OSS), but for brevity
this paper uses OSS.
Often cited advantages of OSS are low cost and freedom from vendor lock-in.
Because few libraries use open source ILSs, little empirical data exists about their use.
Whether open source ILSs cost less than proprietary ILSs is an open question. OSS is
generally thought to cost less than proprietary software, but the impression might be
misleading. Reported disadvantages of OSS are less ease of use and more need for
2
technical expertise. The materials costs of OSS are low, but the labor costs of OSS
might be higher. On the other hand, freedom from vendor monopoly allows competition
among service contractors, which helps to keep labor prices down. Other reported
strengths of OSS include customizability, portability, and security. Free and open
source software are the most permissive types of software for users. By contrast,
proprietary software is the most permissive for software owners, as shown in Figure 1.
IQE§!L . . . ' Software Free 1 Source Public Proprietary) Patented Category (Software (Software Domain Freeware Shareware Software [Software
Example j ®
&
Permissive to Users
Restrictive to Owners
& ^
<V
if ^
# & if
(6 i?
* .¥
J? # ^ ^
$
/ 1 if #
Restrictive to Users
Permissive to Owners
Figure 1. Software classified by category of intellectual property.
OSS has no licensing fees and is not protected by a traditional copyright
restriction. A library district using an open source ILS would pay the same materials
cost for several branch libraries as it would pay for one library. In contrast, a library
district using a proprietary ILS might pay per-site fees based on the number of libraries
or computers. A common license for OSS is the General Public License (GPL). Under
this contract, which gives permission to copy, distribute, and modify a computer
program, the software and its derivative works cannot be patented or made proprietary.
A fee can be charged for distribution or technical support of OSS, yet a free version of
OSS is usually available by download or compact disc.
Libraries use computers to efficiently acquire, organize, retrieve, and disseminate
information. OSS has been developed to facilitate many library-related tasks such as
cataloging photographs and managing bibliographies. An ILS is a database of a library
collection and patron accounts, and the procedures acting on it. It may contain
functional modules for cataloging, circulation, serials, acquisitions, statistical reports,
patron accounts, and an online public access catalog (OPAC, i.e., catalog search engine).
Integrated refers to the ability of the system to share data among its modules. For
example, the information to order a book may be entered in the acquisitions module,
which may be used by the cataloging module, and searched via the OP AC. This
integration reduces redundant data and effort. A synonym for ILS is library
management system (LMS).
ILSs vary by factors including scalability, database type, operating system
compatibility, support for machine-readable catalog (MARC) record formats, and
interoperability with other library networks and articles databases. These factors can be
influenced by whether an ILS is open source or proprietary. Pure open source systems
rest on open source platforms and components. Therefore, some open source ILSs run
only on the Linux operating system, but not on Windows. The open source database
MySQL is used in several open source ILSs. MySQL was originally smaller and slower
than the proprietary database Oracle. However, its use storing metadata in YouTube,
4
which has served 100 million videos per day since July 2006, has put doubts about its
scalability and performance to rest (Do, 2007; Hoff, 2008). Proprietary ILSs have a
well-established market and outpace open source ILSs in many features such as catalog
integration with online journals and multimedia.
Open source ILSs include Avanti MicroLCS, Emilda, Evergreen, Gnuteca, Koha,
OpenBiblio, PhpMyLibrary, and PhpMyBibli (Corrado, 2004). However, most libraries
use proprietary ILSs such as Innovative Interfaces Millennium and SirsiDynix Horizon
(Breeding, 2007a; Pace, 2005).
Problem to be Investigated
This study will investigate why some libraries have chosen an open source ILS
and will compare open and closed source ILSs on costs and benefits. The potential
benefits of an ILS include specific advantages of open and closed source software as
well as general benefits of all software. Table 1 lists the benefits of each type.
Table 1.
Software Benefits
Open Source
Affordability Customizability Portability No vendor lock-in
Proprietary
Turn-key Advanced features Scalability
General
Reliability Performance Ease of use Support Documentation Interoperability Security
An objective of this study is to discover what priority libraries give to each of these
benefits and how well ILSs meet libraries' needs on them. The study will also analyze
5
ILS costs in terms of initial and perpetual costs, and in terms of materials and labor.
The hypothesis of this study is that an open source ILS is more cost-effective
than a proprietary ILS. In other words, the null hypothesis to disprove is that an open
source and proprietary ILS are the same in terms of cost-effectiveness.
Significance of Problem and Justification
Libraries are concerned about cost because they often have more demands than
resources, they are funded by their communities, and they have a responsibility to
manage community funds wisely. If libraries can serve their communities' needs at a
lower cost by using OSS rather than proprietary software, it would be in their interest to
adopt it. Minimally funded libraries, such as small libraries and libraries in rural areas
and developing countries, may find OSS affordable. Library decision makers are
concerned about the total cost of ownership of an ILS. This includes costs of installation,
documentation, training, support, and maintenance.
After an ILS is selected, a library needs to protect its investment. If the vendor
of a proprietary ILS goes out of business or discontinues support of a product, the library
loses its investment of time and funds in the ILS, must choose a new ILS, retrain staff,
and transfer catalog records to the new ILS. In contrast, a library using an open source
ILS may not have to do any of these things. If the support firm of an open source ILS
ceases operation, the library must choose another source of support, but may have more
than one option, including internal support. By using an open source ILS, a library may
reduce its dependence on a vendor.
To meet the changing needs of users and to make use of new technologies, a
6
library may wish to customize or upgrade its ILS. It is unknown how much significance
libraries attach to customizability. If an ILS is proprietary, a library is prohibited from
changing it and must ask the vendor to change it. If an ILS is open source, a library may
change it or outsource the changes to a contractor of their choice.
The next section describes the history of OSS and its adoption in libraries. The
timeline ends with the current status of Koha and Evergreen, the most widely used open
source ILSs. Following the history is a discussion of the larger context of open
information models, including open source and open access. The last part of the paper
reports a survey experiment on open source ILSs.
7
Literature Review
History
"Software sharing is as old as computers" (Stallman, 2002, p. 8). Early
computers were developed in research laboratories, where they were the subject of
academic study. Computer scientists stored programs and data in electronic, digital form.
As computer storage technology evolved from punched cards to tape and discs, digital
media became easier to copy and share. Researchers in academic and scientific
institutions shared software before it became commercialized and proprietary. This
history chronicles how the open source model started, grew, and came to be used in
libraries.
Early examples of open source software and archives are Netlib and Ibiblio.
Netlib is a collection of mathematical software accessible since 1984 on the Internet
(Dongarra, Golub, Grosse, Moler, & Moore, 2006). Ibiblio, formerly known as
SunSITE, founded in 1992, is an archive of OSS hosted by the University of North
Carolina, now mirrored by sites on six continents. This repository also hosts other open
media collections: music archives, text databases, and special exhibits (Ibiblio, n.d.).
When computing passed from early research to commercial uses, many software
producers restricted copying of software and prevented access to source code. The
prohibition of software exchange sometimes led programmers to circumvent it by
reverse engineering binary code. Proprietary vendors hid source code and profited by
charging license fees. Richard Stallman, an early proponent of free software, stood
against this trend. He was a systems programmer at MIT in 1982 when the laboratory he
8
worked in replaced its shared operating system with a proprietary system. He described
the new situation as follows: "The first step in using a computer was to promise not to
help your neighbor. A cooperating community was forbidden. The rule made by the
owners of proprietary software was, 'If you share with your neighbor, you are a pirate. If
you want any changes, beg us to make them'" (Stallman, 2002, p. 9). Stallman
challenged this policy. He declared that "Computer users should be free to modify
programs to fit their needs, and free to share software, because helping other people is
the basis of society" (p. 11).
One of Stallman's formative professional experiences was an encounter with a
Xerox printer that continually jammed. When he wished to modify the printer driver,
Xerox would not release the source code. He could not use a friend's copy, because the
friend had signed a nondisclosure agreement (Bretthauer, 2002). Frustration ensued
because the technical capabilities of the computers owned by his organization were
stymied for nontechnical reasons. In 1985 Stallman founded the Free Software
Foundation. Its object was to create free software, including a free operating system,
GNU (a recursive acronym for "GNU's not Unix").
An early operating system, Unix, began as a proprietary system and eventually
spawned an open source version, Linux. A multi-user system that existed before
Microsoft Windows, Unix was originally developed by Ken Thompson and Dennis
Ritchie at AT&T Bell Laboratories in the 1960s and 70s. Development of Unix
continued at UC-Berkeley resulting in the Berkeley Standard Distribution. Unix was
widely used in universities by computer science students and is many programmers'
9
operating system of choice, due to its transparency and powerful functionality for skilled
users. However, when students left the university, they no longer had access to their
familiar development platform. In 1991 a student at the University of Helsinki, Linus
Torvalds, developed Linux, which was a Unix operating system built on his personal
computer from scratch, free of proprietary sources.
The coincidence of the GNU project needing a kernel (operating system nucleus)
and Torvalds finishing his project near the same time resulted in his contribution of the
Linux kernel to the GNU system. The combination of GNU and Linux was a software
platform, free from the ground up, with no layers of proprietary software hindering
redistribution or modification. It was the first open source version of Unix that could be
run on personal computers.
Berkeley also contributed source code by releasing its Transmission Control
Protocol/Internet Protocol (TCP/IP) as freely redistributable in 1989 (Bretthauer, 2002).
The TCP/IP protocol runs on all computers that communicate on the Internet, including
Web servers and personal computers. The first major open source product that came to
the attention of nonprogrammers was the Netscape Web browser. In 1997 Netscape
released their browser at no cost for the general public, and a few months later, released
the source code (Open Source Initiative, 2005). After Netscape discontinued
development, the open source community carried on support of this project as the
Mozilla browser.
Linux is one part of a development paradigm known as LAMP, which is a group
of open source components for developing Web and database applications such as
10
library OPACs. LAMP stands for Linux, Apache, MySQL, and PHP/Perl/Python, each
of which is open source. Raymond (2001) believes the Linux operating system is the
most successful example of OSS, due to its name recognition and large developer
community. Today, Linux is estimated to have several million users (Linux Counter,
2005), although it is difficult to know exactly how many free copies have been
downloaded or shared.
Other examples of OSS used in production environments are TCP/IP and Apache.
The Netcraft Web server survey (2007) reported that the Apache Web server, which was
released in 1995, held at least 50% of the market share of Web servers from 1996 to
2007, peaking at 71% in November 2005. In March 2008 the largest shares of the Web
server market were held by Apache with 51% and Microsoft with 36%.
The term open source was coined by Christine Peterson in 1998 (Bretthauer,
2002), and the adopters of the term formed the Open Source Initiative. Part of the free
software community began using the term open source software rather than/ree
software so that^ree as in liberty to copy or modify would not be confused with free as in
free of charge. Another aim was to appeal to and avoid alienating for-profit businesses.
Stallman (1999, p. 69) acknowledged the former reason as a valid goal, but
rejected the for-profit issue, which he felt corrupted the spirit of the free software
movement. He objected that OSS allows the "inclusion of proprietary software and
ignores the philosophical issue of software freedom" (Bretthauer, 2002). Eric Raymond,
a proponent of open source, believed the best aspect of OSS was its superior engineering
quality, owing to the large number of collaborators who could improve the software.
11
Stallman (2002, p. 82) pointed out that makers of proprietary software could use the
same argument of software quality, and he preferred to emphasize the freedoms of free
software.
OSS is compatible with commerce. Raymond (2001) presented business models
for supporting OSS. Rather than sell the software, open source businesses may sell
services such as assembly, installation support, and continuing support contracts while
permitting free use, distribution, and modification of the software. For example, the
publicly traded Red Hat company provides a packaged, precompiled (i.e., ready-to-run)
version of Linux with documentation, training, and service agreements, yet the source
code is subject to the open source license clause allowing free distribution (Red Hat,
2005a, 2005b).
The evolution of commercial open source companies has not been without
controversy. Some commercial firms use the term open source mainly as marketing
cachet, but engage in business practices that are not aligned with open source principles,
such as charging subscription fees in place of license fees for bug fixes, while obscuring
information about locating the source code of the fixes. If most of the development of
an open source project is controlled by a commercial company to serve its largest clients,
the benefit to the original community of users may be lost (Moczar, 2005). Perhaps the
most democratic, community-oriented open source projects are those guided by
nonprofit foundations, which coordinate efforts and advocate for developers and users.
Raymond (2001) emphasized the expansion of open source development from its
beginnings in core computer science tools to a variety of uses in many disciplines. He
12
noted that as software tools become standardized as homogeneous commodities, they are
ripe for open source development. This trend coincides with findings by Pace (2005,
p. 25), who surveyed libraries on their ILSs. He observed the "commodity-like nature of
basic ILS functionality" and the "commodified similarities of a wide range of library
automation systems."
The library community first began to take note of OSS in 1999, when Daniel
Chudnov, founder of the Open Source Systems for Libraries project, wrote an
introductory article in Library Journal. Chudnov mentions three library-related OSS
projects: an interlibrary loan document delivery system at Ohio State University, an
automated linking system to electronic resources at Yale, and a draft of an ILS by
Frumkin at the University of Arizona.
Chudnov (1999) notes similarities between philosophies of librarians and
advocates of OSS. Both seek to share resources and contribute to their communities.
Appealing to their common belief in freedom of information, he cites a quote from
Stallman, "libraries should actively discourage the concealment of generally useful
knowledge, and that includes proprietary software" (p. 41). It is possible to interpret this
as encouraging theft, but since Stallman founded the Free Software Foundation to
provide an alternative to proprietary software, it is likely that he is merely discouraging
the use of proprietary software (Stallman, 2002, p. 58).
Chudnov observes that librarians have been more cautious in their consideration
of OSS than other library issues such as intellectual freedom and disability access. One
possible reason for this reluctance, he speculates, is the perception that "you get what
13
you pay for," in other words, "purchase legitimizes quality." However, there are
certainly counterexamples to this adage. For instance, in February 2008, Netcraft rated
Web hosting companies by fewest failed network requests. Seven of the top ten ran on
open source operating systems (Linux or FreeBSD), and only two on a proprietary
operating system (Microsoft Windows).
To promote OSS, Chudnov (1999) suggested that library system vendors release
their source code and sell service contracts. He urged librarians to take part in shaping
open source library software by helping with evaluation and documentation.
Tennant (2000a, 2000b, 2003), a user services and information systems librarian,
continued the advocacy of OSS for libraries. He encouraged librarians to install OSS on
a trial basis. Tennant asserted that open source is better than proprietary software
because libraries may alter it to meet their needs, and such alterations may benefit other
libraries as well. However, he noted that small libraries were unlikely to have
technically sophisticated personnel who could install and maintain OSS, and large
libraries exceeded the scalability limits of open source ILSs at the time.
Cervone (2003) and Buchanan and Krasnoff (2005) considered advantages and
disadvantages of OSS for libraries. One advantage of OSS is portability. The source
code can be compiled for any combination of hardware and operating system. For
example, Koha (www.koha.org) is supported by vendors for Linux and Windows, and
has an informal, unsupported version for Mac OS X. Disadvantages of open source,
which are especially applicable if the software has a small user base, include a lack of
peer review, timely bug fixes, and support. Despite these, use of open source has grown
14
in libraries, in evidence by the dozen library-related OSS applications that Cervone
mentions.
Since the late 1990s, several open source ILSs have been developed (Boss, 2005).
While open source ILS developers have been struggling to provide basic features,
proprietary ILS vendors surpass them with increasingly complex features such as live
chat reference and video, making it difficult for OSS to enter the ILS market (Breeding,
2002). The most widely used open source ILSs are Koha and Evergreen.
A group of four rural libraries in New Zealand contracted in 1999 to develop
Koha. Libraries were slow to adopt Koha, in part due to its original lack of MARC
support, the standard international format for bibliographic catalog records (Hedges,
2003, 2005). The first public library in Northern America to adopt it funded the
development of MARC compliance and began using Koha in 2003. Koha is now used
worldwide, mainly in small libraries (Breeding, 2007b; Koha Developer Wiki, 2007).
Evergreen (open-ils.org) is an open source ILS that was developed beginning in
2004 by a consortium of public libraries in Georgia. It had been installed in 265 libraries
by the time of this study. The source code is available for download under the GNU
GPL. The consortium believes their ILS will lower costs of software maintenance,
licensing fees, and hardware.
The availability of OSS in libraries today is due to the contributions of librarians,
programmers, volunteers, and sponsoring libraries, many of which are funded by the
public. OSS was originally developed, used, and shared by computer scientists. The
first instances of OSS were technical applications: compilers, operating systems, and
15
networking tools. Enabled by the Internet, open source developers collaborated globally.
Expansion of OSS to libraries was made possible by the commoditization of ILS
functions and standardization of bibliographic record format. Past successes and failures
of open source in technical applications provide a framework of advantages and
disadvantages against which to evaluate open source ILSs.
Theory Identified as Framework for Research
The potential advantages and disadvantages of library OSS form the basis for
inquiry. This study will investigate how these factors affect libraries' selection and use
of an open source ILS. Potential advantages of OSS are:
• Lower costs
• Ease of licensing restrictions • Customizability • Portability • Code peer review (if a large enough user base) • Opportunity to contribute to ILS code base • Flexibility in choosing a service contractor
Potential disadvantages are:
• Need for technological sophistication • Higher labor costs • Lack of scalability
• Fewer advanced features
Open source projects with large communities benefit more than small ones from
code peer review. For example, the Linux community has produced many manuals, bug
fixes, software upgrades, Web discussion forums, install fest events, etc. However, open
source ILSs have smaller communities and fewer resources to keep up with demands for
bug fixes and new features.
16
Raymond (2001) warns against the dangers of proprietary software vendor lock-
in. A user of proprietary software has only one source for bug fixes, feature
enhancements, or adaptation to new platforms as technology changes. His investment,
including costs of new equipment, staff training, and data migration, is lost if the vendor
does not perform.
Security is important to software users. Confidential data in libraries include
patron contact information and circulation records. It is unknown but much debated
whether open or closed source software is more secure. Vendors of proprietary software
claim that hiding the source code keeps it safe, whereas revealing the code leaves it
vulnerable to crackers (i.e., those who would exploit the software to cause computer
viruses). Open source proponents counterclaim that many independent developers
search for security holes to plug them, and that the more eyes examining the code, the
more quickly vulnerabilities are found and fixed. Comparisons of the security of open
and closed source software have been inconclusive.
While paid technical support may be available for OSS, a benefit of OSS is the
free support found in Internet mailing lists and discussion forums. A paradox exists
about technical support and why open source ILSs are not quickly adopted by libraries.
Libraries hesitate to adopt OSS if technical support is inadequate, but technical support
is likely to remain sparse until more libraries adopt OSS.
Open Information Models
Open source is one of many open information models, including open standards,
open access publishing, open data, and open courseware. Although seen as alternatives
17
to established means of information sharing, these models are based on values similar to
traditional principles of librarianship and copyright law. Both librarianship and open
source have met challenges to freedom of information and access. In particular, open
source faces a threat from patents. This section will look at what open information
models have in common with traditional models of access and distribution, and will
discuss challenges common to libraries and open information models.
The purpose of open information models is to increase access and decrease cost.
Open information models share values with freedom of information laws, public
broadcast media, the American Library Association (ALA) code of ethics, and the
original purpose of copyright law in the U.S. constitution. Barriers to information access
are often due to economic conflicts with distributors who put profit before public access.
Traditional principles and open information models. The ALA Code of Ethics
(2008) sets forth values of equity of access and intellectual freedom. Chudnov (1999)
observes that libraries and open source serve their communities in similar ways. Both
are public works that make materials available collectively beyond what people could
afford individually.
The ideals of the open source community are similar to those of Ranganathan, a
librarian whose contributions to librarianship included promoting public libraries in
India. He said, "the vital principle of the library ... is that it... freely distributes all the
tools of education and disseminates knowledge with their aid" (1931, p. 415).
Ranganathan's renowned five laws of librarianship express values central to the library
profession. Cana (2003) analyzed the congruence of open source with Ranganathan's
18
laws and wrote laws of open source that parallel Ranganathan's such as "Software is for
use." Gorman (1995) offered an update of Ranganathan's laws. His modernized
restatement could also be applied to open source: "... serve humanity," "protect free
access to knowledge," and "use technology intelligently to enhance service."
Developers and users share and improve OSS for the benefit of society. As a
vehicle for collaboration, the idea of open source agrees with the original intent of the
clause in the U.S. constitution to "promote the progress of science and the useful arts"
(U.S. Const., art. I, § 8, cl. 8). It also accords with the initial aim of intellectual property
(IP) law, which was to encourage people to "build freely upon the ideas and information
conveyed by a work" rather than to "reward the labor of authors" {Feist Publications, Inc.
v. Rural Telephone Service Co., 1991, f̂ 19). IP law affects the sharing of ideas and has
been the subject of controversy concerning patents and open source.
When the U.S. constitution was written, the government conceded patents for a
limited time monopoly (now about 20 years) to inventors as an economic incentive.
Jefferson, the first patent system administrator, felt that it was not a natural right of
people to own an idea, and he was therefore reluctant to grant patents. An inventor
himself, he never patented any of his ideas, because he felt such monopolies were wrong.
To Jefferson, ideas were part of a flow of give and take: an idea does not spontaneously
arise in a solitary mind. We take ideas from others:
... One new idea leads to another, that to a third, and so on through a course of time until some one, with whom no one of these ideas was original, combines all together, and produces what is justly called a new invention. (Jefferson, 1818/1905, p. 88)
19
And we give ideas to others:
He who lights his taper at mine, receives light without darkening me. That ideas should freely spread from one to another over the globe ... seems to have been peculiarly and benevolently designed by nature, when she made them, like fire... (Jefferson, 1813/1986, Tf 1)
Similarly, Stallman (2002, p. 55) noted that software resembles energy more than
tangible property, because transferring it does not deprive an owner of possession.
Patents are incompatible with open source. In the early years of the U.S. patent
office, patents were granted only for inventions that were leaps of scientific progress.
Before 1980, software was generally not patentable, because it was considered
equivalent to mathematical algorithms. As time went on, court rulings set legal
precedents that loosened patent standards. The scope of patents broadened to encompass
software, business methods, and genes. Gradually the scientific standard of showing an
invention was novel and nonobvious shifted to a legal standard of showing there was no
prior patent. Patent creep led to patent inflation.
Ironically, although a patent application reveals an invention, a patent grant
inhibits use of the invention. Rather than encouraging openness and sharing, patent laws
now further the aims of corporations over the rights of individuals (Torvalds & Diamond,
2001, pp. 210, 214). The patent system has become a battleground of corporate interests
that flood the system by practicing defensive patenting to amass portfolios of patents.
These patents function as bargaining chips for cross-licensing in case of patent
infringement claims. The portfolios are an IP landmine for competitors who stumble
upon a patent by independently reinventing it in the normal course of work. In 2006 the
20
U.S. Patent and Trademark Office reported a record 173,772 patents granted, which is a
21% increase over 2005. Of these, IBM was granted the most patents with 3,651, and
Microsoft was granted 1,463 patents in 2006 (Gruener, 2007).
The patent system has been abused by applications for trivial software patents
such as the not equals sign (Bergstra & Klint, 2005). Bessen and Hunt (2004, p. 47)
estimate that 1,080 software patents were issued in 1980, growing to 24,891 in 2002.
Software patents are incompatible with open source because they exclude everyone
except the inventor from the right to use, make, or sell the software. However, inventors
may give permission to exercise these rights through licenses. The common open source
license GPL requires that any patented software it covers must be licensed for everyone's
royalty-free use.
Software patents pose a challenge to open source developers because patents are
more difficult to overcome than traditional copyrights. Although identical copying
infringes copyright, developers may write code from scratch to avoid copyright
infringement. However, inventing around a patented method requires greater ingenuity
because a patent is tied to the method itself, rather than the expression of it as in
copyright. Software that has become an industry standard may be patented, preventing
its use in OSS. For example, MP3 audio clips may not be used in OSS because this
would infringe on the patent held by Fraunhofer, which expires circa 2011-2017
(TuneQuest, 2007). Until then an alternative open source audio format, Ogg Vorbis, is
available but not as popular.
Excessive IP protections are barriers to intellectual freedom and information
21
access. Although companies may return IP revenue to fund research and development,
some studies cautiously suggest that high rates of patenting hinder scientific progress
(Murray & Stern, 2005; Sampat, 2006). In the life sciences, the royalty fees of patents
on genetic materials increase the cost of medical tests and cures beyond affordability for
researchers and patients. Information stakeholders have urged reform of the patent
system (Federal Trade Commission, 2003; Knuth, 1994). A ruling to tighten patent
standards retroactively (KSR International Co. v. Teleflex Inc. et al., 2007) might signal
that the U.S. Supreme Court has taken notice.
Open standards. Open standards are specifications such as equipment sizes, data
formats, and network protocols. The purpose of open standards is to facilitate
interoperability and communication. Like open source, open standards can be developed
and used without significant price barrier. Ideally, members of open standards
committees represent a financially neutral public interest.
The earliest example of an open standard for libraries was the size of catalog
cards. This decision was made at the first ALA meeting in 1877. It enabled mass
production of catalog cards and cabinets (Coyle, 2002). A modern example of an open
standard used in libraries is the MARC record format. Any vendor may implement the
MARC standard, and there is no fee to use it. A MARC record created by a cataloger in
one library can be read by a patron in any other library that uses MARC-compliant
software, even if the reading software is made by a different vendor. Contrast this with a
proprietary standard such as the Microsoft Word .DOC format. If a user exchanges a
Word document with a user on a non-Windows system, the Word document may not
22
display or print properly. Developers have tried to overcome this incompatibility by
reverse engineering the Word document format, but not always successfully.
Coyle (2002,13) distinguishes between motives for open and closed standards,
"If you need to compete, then openness is a disadvantage. If you need to cooperate, then
openness is the way to go."
Open access. To serve their communities, librarians are challenged to balance
library budgets with increasing journal prices. From 1984 to 2004, journal prices
increased about 500%, or 9.36% per year (Dingley, 2004, p. 8). Meanwhile, the total
materials expenditures of academic libraries increased just 250%) (Association of
Research Libraries, 2008), and the Consumer Price Index increased only 82% (Bureau of
Labor Statistics, 2008). Amid this serials crisis, an alternative that has helped scholars
and libraries is open access publishing.
During the 1960s, the U.S. government increased funding to university research,
thereby increasing the number of researchers and articles published. Scientific journal
publishers such as scholarly societies and university presses, which were mainly
nonprofit, could not keep up with the increased literary output. Hence commercial
publishers entered the academic journals market. Because scholars need to keep abreast
of their fields, the demand for journals is high, and because research articles are unique,
their content has no substitute. Thus price inelasticity is taken advantage of by
commercial publishers to set profit margins for scientific journals as high as 42% (Kirby,
1997; Snoddy, 1997).
Commercial publishers took over the majority of scholarly publishing (House of
23
Commons Science and Technology Committee, 2004, p. 13) and consolidated their firms
by mergers and acquisitions. The resulting small number of large publishers holds a
near monopoly on scholarly publishing, limits competition, and fixes prices artificially
high (Case, 2004; Hane, 2003; Kutz, 2002). The large publishers offer big deals of
bundled electronic journals to libraries. Librarians who subscribe to big deals have
trouble learning the prices of individual journals and lose flexibility in choosing journals.
In response to out-of-control serials prices, academic librarians canceled nonessential
journals, decreased monograph budgets, developed collections collaboratively, and
joined consortia to purchase online full-text access. Before canceling journals, librarians
asked faculty which journals were indispensable and apprised university administrators
that budgets were not enough to maintain service levels. Thus, scholars became aware
of the serials crisis.
Scholars who were authors and referees asked themselves why they provided free
labor to commercial publishers who charged high prices and put profit before
scholarship (Bergstrom, 2001). Scientists signed open letters to boycott high-priced
journals like Elsevier's Cell Press (Walter & Yamamoto, 2003), and the editorial boards
of several journals mutinied to start their own journals. Among them, most of the editors
of the Journal of Academic Librarianship resigned after a subscription price increase by
Elsevier, and some of the editors then created Portal: Libraries and the Academy at
Johns Hopkins University Press (Suber, 2006). Enabled by the Internet, scholars have
reclaimed some control of their literature by containing costs in nonprofit publishing.
One way to regain affordability in journals is open access.
24
Open access is freely available, unrestricted access to scholarly information on
the Internet. It has two main variants: archives and journals. Authors may self-archive
papers on their websites or in long-term repositories of their subject area or institution.
About half of traditional journals allow self-archiving (Gadd, Oppenheim, & Probets,
2003, p. 26; McVeigh, 2004). The second open access variant is publishing in peer-
reviewed open access journals. The costs of open access publishing are borne by the
author or author's institutional sponsor. Open access authors retain copyright, and grant
permission to their articles through licenses such as the Creative Commons license.
Articles in both open access variants may be indexed in search engines using the Open
Archives Initiative Protocol for Metadata Harvesting (OAI-PMH), which enables
searching across multiple repositories.
In 1991 Ginsparg organized a preprint server, ArXiv, to distribute physics
articles freely online. ArXiv was the first centralized open access archive of scientific
articles. Scholars from other disciplines have also submitted preprint articles to the
server, and it has grown to 470,000 articles and 16 mirrors on five continents (ArXiv,
2008a, 2008b). Since 2004, ArXiv has received over 3,000 submissions per month
(ArXiv, 2008c). These articles, also called eprints, are generally not peer-reviewed,
although authors may replace their original submissions with the final form of published
articles accepted by peer-reviewed journals. An advantage of eprints is that they are
circulated earlier than articles in traditional journals. Physics publishers have not
reported a loss of revenue due to the ArXiv (Swan, 2005, p. 7). Schwarz and Kennicutt
(2004, p. 7) note that the destination of many ArXiv eprints to peer-reviewed journals,
25
which have rigorous publication standards and editorial requirements, keeps the quality
high.
Disciplines such as medicine have no preprint culture due to the potential
liability of publishing inaccuracies. Researchers in these fields have sought the peer-
reviewed open access journal alternative. The Public Library of Science and BioMed
Central are examples of open access journal publishers. Some journals are hybrids,
offering authors a choice of traditional or open access publication.
Large commercial publishers, who stood to lose profits, claimed that open access
would destroy the peer review system, deprive authors of copyrights, exclude authors
who could not afford to pay page charges, and bankrupt nonprofit publishers. But the
peer review system, which is compatible with open access, continues to be rigorous.
Authors retain copyright, and license permissions to open access publishers, repositories,
and readers. Page charges may be paid by authors' sponsoring institutions or waived in
case of authors' financial hardship.
The sustainability of business models for open access has not yet been
established. Embargo periods and page charges to authors or sponsors vary with the cost
of producing open access journals. It is uncertain what the effect of shifting journals
funding from libraries to research institutions will be. Among the opposition to open
access, the most compelling concern is that of traditional nonprofit scholarly publishers.
Their journals tend to be high quality publications with low subscription prices, and their
quality to cost ratios may be higher than commercial or open access publishers.
Studies have compared the research impact of open access and traditional journal
26
articles by using citation counts as a proxy for research impact. In 2004 McVeigh
studied data from the Thomson ISI citation database and found that traditional journals
still offered their authors more visibility than open access journals (pp. 5, 16). However,
this study may have been premature because it used data from two years earlier when
few open access journals existed, and search engine indexes of open access articles were
incomplete. By contrast, Antelman (2004) found that open access articles across four
disciplines had a greater research impact than articles not freely available. When
comparisons were made between open access and non-open access articles within the
same journal, studies showed that open access articles were cited two to three times as
often as non-open access articles in the Astrophysical Journal (Schwarz and Kennicutt,
2004) and Proceedings of the National Academy of Sciences (Eysenbach, 2006).
Mandates for public access to publicly funded research. The results of research
sponsored by taxpayers are often patented and copyrighted. This privatization of
knowledge removes the benefit from the public who paid for it. To restore the
investment to the people, legislators have tried to require public access to publicly
funded research.
Congressman Martin Sabo, in 2003, proposed the first bill to mandate public
access to the results of publicly funded research, titled the Public Access to Science Act
(H.R. 2613). Mr. Sabo proposed to remove copyright protection from federally
sponsored researchers, as is done to federal employees. The proposal sparked
controversy, and the bill did not emerge from its congressional subcommittee.
Three weeks after a Harris Poll (2006) showed that 82% of U.S. adults were in
27
favor of making results of publicly funded research freely available online, Senators
Cornyn and Lieberman introduced the Federal Research Public Access Act of 2006
(S. 2695), with the same purpose as Sabo's bill but leaving authors' copyrights intact.
Even though it had bipartisan sponsorship, it suffered the same fate as the Sabo bill
because of opposition from publishers.
A 2006 report of the European Commission recommended "public access to
publicly funded research results shortly after publication" (recommendation Al, p. 11).
In April 2007, the European Commission incorporated this open access mandate into its
scientific grant application (Seventh Research Framework Programme, 2007). The
Wellcome Trust (2008), a sponsor of medical research in the United Kingdom, has
required grantees to deposit publications since October 1, 2006, into the permanent, open
access archive PubMedCentral within six months of publication. The U.S. National
Institutes of Health followed suit with an open access mandate (with a maximum of a
one year embargo) effective April 7, 2008. The National Science Foundation has also
resolved to make citations to research it funds publicly accessible online (NSF Office of
Inspector General, 2006, p. 8).
Summary. Open information models provide users equitable access to a
commonwealth of knowledge. Corrado (2005) states that the benefits to libraries of
open source, open access, and open standards include increased access, lower costs, and
better long-term preservation of scholarly works. Large commercial publishers and
software corporations oppose open information models because they may lose profit.
Due to steep increases in journals prices, scholars and librarians are switching to lower
28
cost and open access journals where possible.
Willinsky (2005) believes that by recognizing the similarities of different types
of open initiatives, information stakeholders can unite to make information and
technology accessible and economical. He asserts that using OSS to help in the
operations of open access journals is a powerful combination. In order to prevent patent
misuse, which may hinder scientific progress, scientists have also begun to consider how
open source collaboration can be applied to the life sciences to share techniques and
materials.
Current Research Findings and Gaps
Chudnov (1999) noted that libraries have not been quick to adopt OSS. Pace's
(2005) and Breeding's (2007a) data on the number of sites per ILS vendor show that
most ILSs in use are proprietary, and fewer than two percent of ILSs in use are open
source. Few open source ILS products exist, they are not well known, and their software
is not as mature and feature-rich as proprietary ILSs. By 2007 among public libraries in
North America, only one state consortium (Georgia) of 265 libraries, two county library
systems with about seven branches each (Athens County, OH, and Crawford County,
PA), and four individual libraries were using open source ILSs.
Buchanan and Krasnoff (2005) observe that, in the school library environment, a
completely open source solution might be impractical. Much educational software runs
only on proprietary platforms, such as Windows or Mac OS. If an ILS were available
only for an open source platform, this might necessitate purchase of new hardware, thus
lessening any cost savings.
29
Most articles about library OSS are advocacy essays and many describe library
applications other than ILSs. Clarke's (2000) master's thesis described the history of
OSS in the library community, related the details of several library OSS projects and
suggested that library OSS be studied further when it matures. Proffitt (2002) reported
on the use of Koha at the Athens county library district in Ohio, and Cargile (2005)
narrated her attempt to install an open source ILS. Among articles that are directed at
librarians, Cargile observed a scarcity of practical information about OSS, such as
simplified technical explanations.
A few articles describe empirical studies of open source ILSs. Kumar (2005) has
compared the open source ILSs Koha, PhpMyLibrary, and OpenBiblio in a cross-
comparison ranking of their features, and found that Koha was the most functionally
mature of the three. Barbier (2005), Berizzi and Zweifel (2005), and Chalon, Alexandre-
Joaquim, Naget, and Becquart (2006) recounted their decision-making processes in
selecting an open source ILS as follows.
Barbier (2005) compared Koha with PhpMyBibli (PMB) on factors such as ease
of installation and use, import and export of records, robustness, and security. Although
he found PMB's interface more user-friendly, Barbier recommended Koha partly
because it handled different MARC formats (MARC21 and UNIMARC). Barbier also
questioned PMB's future as OSS because it is backed by a commercial firm. Berizzi and
Zweifel (2005), and Chalon et al. (2006) researched several open source ILSs for small
collections. After unsuccessfully trying to install Koha, both settled on PMB. Chalon et
al. sought help in the forums and mailing lists. They felt that questions about installation,
30
hosting, and migration may go unanswered because of conflicts of interest with
commercial open source firms that provide these services.
To choose an ILS for a group of libraries in the UK, Wilson (2005) surveyed 18
library staff and 4 patrons on three proprietary ILSs and one open source ILS. Many of
the respondents tested all of the systems on the same day and were able to compare them
directly. The respondents scored the ILSs on functional modules and common tasks.
Respondents commented that Koha may have been "at a disadvantage because [it was]
not presented as a sales pitch" and it was "Basic but functional. Clearly built to cost"
(p. 5). Although Koha was rated lowest in every category behind FDI OLIB7, Sirsi
Horizon, and Innovative Millennium, the group of libraries ultimately chose it. Lastly, a
survey of approximately 1,000 Evergreen patrons (Georgia Public Library Service
[GPLS], 2007) found that most respondents agreed or strongly agreed with positive
statements about Evergreen, and many patrons found the Evergreen self-service features,
such as placing online holds and renewals, easy to use.
Koha
Development of Koha, an open source ILS, began in 1999, funded by a group of
libraries in rural New Zealand that found proprietary software expensive and lacking in
needed features (Breeding, 2002; Proffitt, 2002). According to Blake (2000), Koha
started out as a "rush job" to replace a system that wasn't Y2K compliant. The libraries
licensed the ILS under the GNU GPL to make Koha freely available to other libraries
and to avoid tying themselves to a particular software firm for maintenance.
Koha is written in the LAMP paradigm. Early adopters of Koha were
31
programmers who were in some way connected with libraries. Shortly after Koha was
released for public download, developers who had independently written library
software contributed new features to Koha. In 2002 only two school districts in North
America were using it. A lack of compliance with interoperability standards limited the
usefulness of the early open source ILS.
The first public library in the United States to use Koha was the Nelsonville
Public Library, which serves Athens County, Ohio. They chose Koha because they were
frustrated that they could not customize their proprietary ILS to their needs. The library
financed the development of Koha's support for the MARC record format and Z39.50
information retrieval protocol. Z39.50 is an Internet communications protocol used in
copy cataloging to search and transfer MARC records. This feature, which was more
highly developed in proprietary ILSs, enabled interoperability with other systems, such
as retrieving bibliographic catalog records from the Library of Congress. Henceforth,
the added Z39.50 and MARC features and their source code were available to any
library.
Koha went live in Athens County in 2003 in a district of seven library branches,
60,000 patrons, and a collection of 250,000 items (Hedges, 2003, 2005). The main
library in Nelsonville spent several thousand dollars training staff in the MySQL
database and programming languages PHP and Perl. The library director said the initial
expenditure to add desired features was less than the annual fees of proprietary
alternatives, "We may be spending $10,000 this year, but we will be saving $10,000 in
license fees for every year thereafter" (Proffitt, 2002, p. 2).
32
The Koha source code shows that many different authors retain copyright to their
programming contributions. Anyone may contribute to the code, but most development
is currently done by a commercial open source firm begun by former employees of the
Nelsonville library. The company, Liblime, has bought the Koha Web domain name,
acquired the division of the New Zealand company that originated Koha, and hired
employees from major proprietary ILS firms. An elected release manager who controls
the software development is currently employed by Liblime. It is unknown whether the
concentration of Koha resources in a company whose focus is "work for hire" rather
than Koha is best for the Koha library community.
A turnkey system for Koha with technical support became commercially
available in 2006, which could relieve libraries of installation and maintenance tasks.
Koha has increased its scalability and performance by optimization for indexing
(Liblime, 2005). Over 250 libraries worldwide use Koha (Breeding, 2007b; Koha
Developer Wiki, 2007).
Evergreen
Evergreen is an open source ILS developed beginning in 2004 by a consortium of
public libraries in Georgia. Currently the largest open source ILS, it was installed in
2006 in 252 libraries, increasing to 265 at the time of this survey. Scalability has not
been a limitation of this ILS, which accommodates 1.8 million patrons and a collection
of 9 million items (D. J. Very, personal communication, March 4, 2008). Evergreen
uses clusters of Linux servers, an arrangement similar to that used by Google, which
provide load balancing for performance, and redundancy for reliability (LaJeunesse,
33
2005). The underlying database is PostgreSQL (GPLS, 2006b), and the staff client
software runs on Windows, with Linux and Mac support planned (LaJeunesse, 2006).
Patrons access their accounts securely through a Web OP AC.
In the future, linkage with full-text article databases is planned (GPLS, 2006a).
In 2007 the Michigan Libraries Consortium announced plans to install Evergreen
(Michigan Evergreen, 2007). British Columbia, which installed Evergreen in three
libraries, estimated the cost of the open source ILS as one fifth of a comparable
proprietary ILS, a savings of $8 million Canadian dollars over the next five years
(British Columbia Public Information Network for Electronic Services [BC PINES],
2007). Development is carried out by the Equinox company, started by former GPLS
employees. Work on serials and acquisitions modules is being contributed by several
Canadian universities.
34
Research Procedures
Survey Introduction
The purpose of this survey is to compare costs and benefits of open source and
proprietary ILSs, and to find out the factors that influence a library's decision to adopt an
open source ILS. To determine the ILS costs, questions were asked about annual cost,
initial cost, and related cost of required system components. Questions about benefits
took the form of satisfaction ratings on 5-point Likert scales. Additionally the
questionnaire asked the reasons for ILS selection, whether the source code was
customized, who provided system services, where the ILS was hosted, what problems
were encountered, and library size and type.
The open source ILSs Koha and Evergreen were targeted in the survey. Other
open source ILSs exist, but Koha and Evergreen were chosen for this survey because
they had the largest user communities. At the time of this survey, which was conducted
on October 30, 2007 to January 3, 2008, the current version of Koha was 2.2.9 and
Evergreen was 1.2. The survey was answered by respondents using Koha (n=l 13),
Evergreen (n=l 19), and proprietary ILSs (n-129). The survey respondents represent
about 40% of known Koha libraries, and 30% of Evergreen libraries, but less than 1% of
libraries using proprietary ILSs. The respondents were mainly librarians and IT
personnel such as system administrators and computer programmers. About two each of
the Koha respondents were consultants, students, or had installed Koha in home libraries.
Koha and proprietary ILS respondents answered the main questionnaire (see
Appendix A). Evergreen respondents answered a subset of the questionnaire, which did
35
not include questions about ILS choice and administration. These questions were
answered for Evergreen at the consortium level by the PINES program director because
the selection, support, and costs were coordinated by the consortium.
Sample and Recruitment
A list of libraries using Koha was prepared from the Koha users list, Marshall
Breeding's lib-web-cats database, email addresses in the archives of English and French
Koha listservs, and Google searches of Koha OPAC URLs (inurlropac-main.pl or
inurkopac-search.pl). This list contained email addresses of approximately 250 Koha
library organizations, some of which had multiple branches. Many candidates for the
Koha survey responded that they were unable to take the survey because they were still
in the setup process, or had not used it very long.
The libraries using Evergreen are part of a voluntary consortium, PINES, in
Georgia. Evergreen sample was prepared from a list of Georgia PINES libraries, the lib-
web-cats database, and an internal Georgia PINES listserv. The total of Georgia PINES
Evergreen (main and branch) libraries at the time of the survey was 265. Although
several libraries in Canada had begun using Evergreen, they were not yet ready to be
surveyed, as they had been using the software less than six months.
The respondents from libraries using proprietary ILSs were drawn from listservs.
Several of the listservs had a particular library type focus: public, academic, special, or
K-12 school libraries. Some of the listservs were international and others were mainly
U.S.-based.
Recruitment of respondents was done by sending invitations to take the survey.
36
The letter to participants described the survey as a cost-benefit study of ILSs, but did not
specifically mention the purpose of comparison between open source and proprietary
systems. This was omitted to avoid exaggerated responses from participants who might
wish to strongly promote one type of system over the other. The invitations, which were
translated into French, Spanish, Italian, Portuguese, and Polish, were emailed to listservs
and individual email addresses. The recruiting was voluntary and involved self-selection,
and thus introduced some bias into the data which lessened the experiment validity.
The proprietary ILS respondents were recruited from listservs only; no individual
recruitment emails were sent to them. Because answering a listserv post requires greater
initiative than answering an individually addressed email, the self-selection bias is
greater in the proprietary ILS sample. Self-selection can draw respondents with greater
interest, knowledge, or stronger views, either positive or negative. One possible reason
for the interest of some of the proprietary ILS users was their looking for a new ILS,
mentioned by 12 respondents.
Another source of bias in the open source sample is clustering. The Evergreen
users belonged to the same consortium, and some to the same libraries or regional
systems. Eleven of the libraries using Koha had more than one respondent belonging to
the same organization. Clustering may cause less variation in the results.
Survey Method
A limitation of this survey is that the benefits of the ILS, as given by satisfaction
ratings, are subjective. Subjective ratings may reflect the rater more than the thing rated.
For example, two respondents may report equal satisfaction with systems whose
37
functionality differs greatly. Another difficulty with soft data is that it may be reported
differently by a respondent for the same question on different occasions. One question
in this survey was asked twice in different parts of the survey. The answers were not
exactly the same, but not significantly different.
Several other problems limited the effectiveness of the survey. A general
problem with web surveys is the possibility of respondents' answering the survey more
than once. Although the questionnaire was reviewed by several people, no pretest was
done. The questionnaire was too long. The length of the questionnaire is a tradeoff
between complete coverage of the topic and the time a respondent can spend on it. To
avoid further lengthening, many terms in the survey were not defined, and the lack of
definitions confused some respondents. For example, many respondents were unsure of
whether their customizations were source code modifications, and thus a follow-up
question was added for clarification. Also, respondents interpreted some terms
differently depending on whether they were librarians or computer technologists. For
example, maintenance was interpreted by a librarian to mean bibliographic cataloging,
but interpreted by a system administrator to mean tasks such as file backup.
Although the survey was too long, lack of detail in the question text caused
inconsistencies in the data. For instance, the library population may have been reported
as the number of patrons who frequently use the library, whose names are registered in
the ILS, or who live in the area. Also, the prerequisite system requirements cost may
have been reported inconsistently: some respondents may have omitted the cost of
hardware they already had. A possibly more accurate survey method is a telephone
38
survey, because it allows for immediate clarification and more consistent recording of
data, but it is more expensive than a web survey.
Respondent Characteristics
If this survey experiment had perfect validity, no factor besides whether the ILS
was open source or proprietary would affect the outcome. But the two main sample
groups—open source and proprietary ILS libraries—differed on other characteristics.
Tables 2-5 show characteristics of the open source and proprietary ILS respondents that
might influence the survey outcomes.
The libraries using open source ILSs in this sample are mostly smaller than those
using proprietary ILSs (Table 2). This is partly because Koha has not been tested in a
large environment. The largest Koha installations answering this survey were two
groups of 20 libraries each.
Table 2
Library Size
Library Library
collection size population size
n
106 104
Koha
mean
46,783 5,823
Open
median
8,750 520
Source
n
265 265
Evergreen"
mean
33,843 6,945
median
21,981 2, 910
n
110 112
Proprietary
mean median
468,532 90,500 260,764 14,500
a The Evergreen data for print volumes and registered patrons as of October 2007 were provided by Diana Very, Director of Statistics and Research, GPLS.
Table 3 shows the library types of the respondents. The libraries using
Evergreen were all public libraries. Public libraries made up 30% of the proprietary ILS
sample group, but only 12% of the Koha group.
39
Table 3
Library Type
Q 55) What type is your library?
Total
Public
Academic
Special
School
Home
Don't know/no answer
Open Source Pro-
Sub Ever- prie Other Total total Koha green tary /na
365 232 113 119 129 4 100.0 100.0 100.0 100.0 100.0 100.0
172 133 14 119 39 47.1 57.3 12.4 100.0 30.2
71 37 37 19.5 15.9 32.7
65 35 35 17.8 15.1 31.0
27 17 17 7.4 7.3 15.0
2 2 2 0.5 0.9 1.8
7.7 3.4 7.1
32 2 24.8 50.0
28 2 21.7 50.0
10 7.8
20 15.5
Table 4 shows the location of respondents by continent. The Evergreen
respondents are all in Georgia, the proprietary ILS respondents are mainly in North
America, and the Koha respondents are the most geographically dispersed. The
proprietary ILS sample might least represent the proportions of type and geographic
region in the actual population because of self-selection bias and the small sample size
relative to the large total population of libraries using proprietary ILSs.
40
Table 4
Library Location
Q 56) Where is your library located?
Open Source Pro-
Sub Ever- prie Other Total total Koha green tary /na
Total 365 232 113 119 129 4 100.0 100.0 100.0 100.0 100.0 100.0
Africa
Asia
Europe
North America
Oceania
South America
Don't know/no answer
7 1.9
20 5.5
32 8.8
244 66.8
17 4.7
8 2.2
37 10.1
6 2.6
11 4.7
23 9.9
165 71.1
12 5.2
7 3.0
8 3.4
6 5.3
11 9.7
23 20.4
46 40.7
12 10.6
7 6.2
8 7.1
119 100.0
1 0.8
6 4.7
9 7.0
79 61.2
5 3.9
1 0.8
28 21.7
3 75.0
1 25.0
As shown in Table 5, Koha libraries, more often than proprietary ILS libraries,
tend to host their ILS (84% vs. 70%) and provide ILS-related services internally (61 % vs.
31%). The Evergreen libraries of the Georgia PINES consortium host their systems
inhouse and originally provided all services for their ILS internally. Subsequently, the
employees who developed Evergreen formed a consulting company, now contracted by
Georgia PINES to provide ILS services.
41
Table 5
ILS Hosting and Service Providers
Koha (n=113) Proprietary (n=129)
inte exte inte exte rnal rnal both rnal rnal both
•sting
oviders: Installation Documentation Training Support Maintenance
84%
61% 59% 57% 66% 58% 66%
12%
22% 29% 20% 17% 23% 19%
-
15% 12% 14% 14% 18% 16%
70%
31% 43% 26% 37% 19% 30%
29%
30% 27% 38% 21% 38% 28%
-
35% 27% 29% 40% 40% 39%
OSS users are at liberty to customize any aspect of the software. Source code
customization means changes written in a programming language such as Perl, but it
does not mean configuring ILS options, changes to OP AC web page appearance made in
HTML, or API scripts to access the catalog database. Libraries using Koha customize
the source code (23%) more often than do libraries using proprietary ILSs (7%). Several
of the proprietary ILS respondents reported source code customizations made by vendors
at the library's request. A few respondents of both proprietary and Koha ILSs noted that
their customizations were lost when the system was upgraded. The Evergreen ILS was
created from scratch in Georgia, thus it could be considered completely customized,
although individual PINES libraries must make requests through the consortium for
further customization. One Evergreen respondent noted, "Almost every suggestion that
'really works' does seem to get integrated into the system eventually."
The differences in the characteristics of the sample groups temper the
interpretation of the findings in the next section.
42
Findings
Reasons for ILS Choice
The Koha respondents in this survey based their choice of ILS on cost before
functionality, whereas the proprietary ILS respondents reversed these priorities.
Libraries using open source ILSs also cited customizability, freedom from vendor lock-
in, and portability as reasons for choosing their ILS more often than libraries using
proprietary ILSs. By contrast, libraries using proprietary ILSs more often cited ease of
use, good technical support, trusted brand name, and having used the ILS before. This
question was not asked of individual Evergreen libraries, but of only one representative
of the Evergreen consortium.
Table 6
Top Five Reasons for ILS Choice
Koha Evergreen Proprietary
1 Affordability Scalability Functionality 2 Functionality Customizability Affordability 3 Customizability Security Ease of use 4 Freedom from vendor lock-in Technical support Technical support 5 Ease of use Ease of use Trusted brand name
Among those who answered whether the selector had analyzed costs and benefits,
70% of open source respondents said they chose their ILS because it was the most cost-
effective, compared with 46% of proprietary ILS respondents.
Eighteen respondents specifically mentioned open source as a reason for
choosing their ILS. Seven open source respondents mentioned social and political
reasons such as "to promote cooperative work," "a different way to spend public funds,"
and "to share with other institutions." Five reported mandates or strong support for
43
using open source at the organizational, regional, or national level. Two respondents
said that long-distance technical support calls were difficult in remote regions with
expensive and unreliable telephone networks; hence they chose open source or
homemade systems. Nine proprietary respondents said their selection was based on a
consortial decision, and five said their ILS was the result of an ILS company buyout
more than a choice on the library's part. Two proprietary ILS respondents also
mentioned that they were concerned about their vendors' future existence.
In the recruitment phase of the survey, about 30 candidates had seriously
considered Koha but chose another solution. Twenty found Koha too complex for their
needs or too difficult to install. Of these, three found inhouse solutions, five chose
proprietary ILSs, and five chose smaller open source ILSs: PMB (two) and OpenBiblio
(three). One person chose Emilda, another small open source ILS, and encountered a
problem with very small OSS projects: the project is no longer supported by anyone.
On the other hand, five candidates found Koha too small, lacking scalability, support, or
an academic reserves module. Two were waiting for the next version of Koha to be
released. One candidate had considered Evergreen but did not adopt it due to its lack of
acquisitions and serials module, and uncertainty about whether the consulting firm had
enough people to provide reliable 24/7 support.
ILS Satisfaction ratings
Respondents were asked to rate their satisfaction with their ILS on a 5-point
Likert scale. Table Bl in Appendix B shows the average satisfaction ratings of various
ILS aspects, the number of responses, difference of satisfaction ratings between the
44
sample groups (e.g., open source vs. proprietary), and two-tailedp-values for the
differences. The table is sorted by the difference (^-statistic) of open source and
proprietary ILSs from the comparatively higher-rated aspects of proprietary ILSs at the
top to the lower-rated aspects of proprietary ILSs at the bottom. A few of the items were
not rated by individual Evergreen libraries, but only at the consortium level. These
items were installation smoothness, acceptableness of time to install, maintenance,
scalability, and affordability.
Overall, the satisfaction of open source ILS respondents is slightly higher than
that of proprietary ILS respondents (4.2 vs. 3.9 on a 5-point scale). The difference is
statistically significant (p < 5%, ̂ -values are two-tailed throughout). However, it is
uncertain whether the difference is due to the ILS being open source or proprietary,
because the samples compared are different in terms of other population characteristics
as shown in the Tables 2-5. This uncertainty also applies to interpreting the rest of the
results in this survey where statistical significance is conjectured.
Installation smoothness, a question that was not asked of individual Evergreen
libraries, was the category in which proprietary ILSs most exceeded the open source ILS
Koha (3.7 vs. 2.9 on a 5-point scale,p < 0.1%). Other categories in which proprietary
ILSs exceeded open source ILSs, from highest to lowest difference, were documentation
completeness (p < 0.1%), acceptableness of installation time—not asked of individual
Evergreen libraries (p < 1%)—and original cataloging ease of use (p < 5%).
At the opposite extreme, affordability, also not asked of individual Evergreen
libraries, was the category in which Koha most exceeded proprietary ILSs (4.6 vs. 3.6 on
45
a 5-point Likert scale,/? < 0.1%). Other categories in which open source ILSs exceeded
proprietary ILSs, from highest to lowest difference, were customizability,
interoperability, portability (p < 0.1%), functionality, OP AC completeness, scalability—
not asked of individual Evergreen libraries, circulation/patron accounts ease of use
(p < 0.5%o), OP AC self-service ease of use (p < 1%>), and overall ILS satisfaction
(p<5%).
The results at the extremes coincide with the literature about open source and
proprietary software. At one extreme, open source ILSs are thought to require more
technical sophistication, and hence present a greater technical challenge to install.
Documentation is also a commonly cited weakness of OSS. At the other extreme, open
source ILS respondents are more satisfied than proprietary ILS respondents with the
affordability and customizability of their systems.
However, as the statistical significance of the difference between open source
and proprietary ILS ratings decreases, some of the other results contradict the literature.
It was unexpected that open source users would rate their satisfaction with functionality,
OP AC completeness, scalability, or ease of use as higher than would proprietary ILS
users.
Three observations can be made about the satisfaction ratings in Table Bl. First,
the most extreme differences were in categories where essentially the only open source
respondents who were asked the questions were Koha users, because installation and
decisions about affordability for Evergreen were handled at the consortium level. Are
these results peculiar to Koha, or inherent to all open source ILSs? If Evergreen is
46
installed in more libraries in the future, more data could be gathered from Evergreen
users to gain insight. Secondly, the proprietary ILSs were rated higher on cataloging
items, while the open source ILSs were rated higher on OP AC items. Perhaps this
reflects that the proprietary ILSs programmers have more bibliographic cataloging
expertise. Third, the highest rated strengths of OSS are more conceptual (affordability,
customizability, interoperability, portability, functionality), while the highest rated
strengths of proprietary ILSs are more concrete (installation smoothness, documentation
completeness, satisfaction with installation time, cataloging ease of use).
Verbatim comments from respondents reflect the numeric ratings. While the
open source ILS respondents' comments about installation and documentation were
occasionally positive, many expressed the need for improvement. Regarding
documentation for open source ILSs, Koha respondents said, "The documentation is not
updated regularly," and, "Certain sections are very complete, whereas others are
incomplete or missing. Plus, depending upon what computer, software version, etc., you
are using, some of the documentation does not match." One Evergreen librarian was not
able to understand the "programming jargon." Some respondents had suggestions for
improving the open source ILS documentation:
• The documentation could be more user-friendly. • It should have more screenshots and be easier to understand, using plain
language. • Needs more attention to getting Z39.50 servers installed and working. • Needs ... language the end user can understand. • More practical details. • Better documentation about MARC fields and setting up parameters.
A preference for documentation written by librarians was expressed by three
47
respondents, "The documentation supplied is written by the developers, but ideally
should be written by users (librarians)." A librarian agreed about writing a manual, "We
need to come forward to do things for [ourselves]," and another added, "As more
librarians use the software, the documentation is getting better." Six open source
respondents said they had or planned to contribute to the documentation. One also
mentioned attempts to bring all Koha documentation into a central Web location.
Another respondent preferred a printed manual. Elizabeth Garcia, the PINES program
director, said the Georgia PINES consortium was getting a grant to write more
documentation for the Evergreen ILS.
Koha respondents expressed difficulties with installation. Respondents said,
"Installation took an inordinate amount of time because of the trial and error of trying to
install on various operating systems..., hardware and software incompatibilities, etc,"
"Too many Perl modules that needed to be downloaded separately," "It took us two and
a half more days to get it to access external Z39.50 servers," and "It is not an out-of-the-
box solution." Four respondents felt that Koha installation smoothness was improving
with time. Three expressed the need for a bundled installer, which would contain all
components and automate the configuration.
ILS Problems
Respondents were asked about problems with ILS modules, data loss, lapses in
service, and security breaches. The users of Koha and proprietary ILSs reported no
significant difference in these types of problems. In an "other write-in" category, six
Koha users mentioned difficulties with the Z39.50 function, which downloads catalog
48
records. The number of problems that Evergreen users attributed to the ILS itself was
not significantly different than that of proprietary ILS users. However, when combining
the problems due to the Evergreen ILS with problems possibly due to another source
such as network outages, the difference was significant (p < 0.1%). Evergreen users also
had significantly more lapses in service (p < 0.1 %), problems with the circulation
module (p < 0.1%), and problems with the OPAC/search engine (p < 0.5%). This may
be due to the recent release of Evergreen, which went live in September 2006, only 15
months before the survey was conducted. By comparison, the users of proprietary ILSs
were using mature software. Eighteen of the proprietary ILS users made reference to the
age of their ILS or said they had weathered ILS company mergers.
Eight Evergreen respondents related the ILS problems to its being relatively new
and still improving. One respondent said that most of the problems were due to the
"software being new and a work in progress," and another said Evergreen was
"undeniably launched prematurely—every staff member became an alpha tester in effect
the first few months." It may have been premature to survey Evergreen users at this time,
and perhaps in a few years the Evergreen problems will decrease. Nine Evergreen
respondents mentioned slow response time of the system. One explained, "There are
many times a day that we have slowness in doing many functions at the circulation desk
and in the cataloging department." Attempts to resolve the ILS slowness have included
adding more Tl network hardware to increase bandwidth and raising the priority of ILS
traffic on the network.
49
ILS Costs
Evergreen costs. Before Evergreen, the Georgia PINES consortium used the
Unicorn ILS made by vendor SirsiDynix, running on a Sun Microsystems server. The
previous costs are shown in Table 7. The old server cost one million dollars initially,
and the annual cost was $558,000:
Table 7
Annual Cost of Previous ILS at Georgia PINES Consortium
Item Cost
• Hardware Sun Microsystems maintenance $56,000 Server/networking costs $102,000
. Software (ILS) License fees $200,000 Service/maintenance $200,000
Total annual cost of previous ILS $558,000
The cost and library size questions about the current Evergreen ILS (Table 8)
were not asked of the individual Evergreen libraries, but rather asked only at the state
consortium level. Data for print volumes and registered patrons as of October 2007 were
provided by Diana Very, Director of Statistics and Research, GPLS. The Evergreen cost
questions were answered by Elizabeth Garcia, PINES program director, in a phone
interview in January 2008.
50
Table 8
Evergreen Size and Costs at Georgia PINES Consortium
Size or Cost
Collection size Service population Consortium size Prerequisite system requirements Annual cost Initial labor cost
Amount
8,968,482 1,840,318
265 $150,000 $450,000 $252,500
Units
print volumes registered patrons libraries
All residents of Georgia, population 9 million, may check out books from the
PINES consortium. The libraries in the Atlanta metropolitan area, which has about half
the state's population, do not belong to the consortium or use the Evergreen ILS.
However, their residents may check out PINES items.
The current cost of $150,000 for prerequisite system hardware of the Evergreen
system includes commodity hardware, networking equipment, and redundant power. The
individual Georgia PINES libraries did not need new hardware to use Evergreen. The
initial (software) materials cost was nothing: all components were OSS. The initial
labor cost of $252,500 consists of:
• $ 15,000 outsourced programming • $ 10,000 employee training and travel
• $227,500 staff programming ($ 130,000/year for 1.75 years)
Ms. Garcia reported that the flexibility of OSS to run on commodity hardware
gives them a great cost savings. When any hardware component fails, the system
administrators can search for the lowest cost replacement, rather than being tied to an
expensive server.
51
The current annual cost of $450,000 for Evergreen at Georgia PINES includes
$350,000 for ILS costs to the Equinox service company and $100,000 for networking
costs. The networking equipment is also used for other non-ILS library functions such
as staff email. The savings over the previous proprietary ILS is more than $100,000
annually.
ILS costs estimated by vendors. To estimate proprietary ILS costs, several
vendors in the United States were contacted for price quotes for a library of 15,000 items.
After removing the highest and lowest estimates, four quotes for initial costs and three
quotes for annual costs were obtained. Some of the estimates did not include
acquisitions or serials modules, or hardware. The mean initial cost was $24,313 and the
mean annual cost was $2,133. Costs may be higher for additional branch libraries. Two
quotes for Application Service Providers (ASPs, which are hosted by the vendor)
conformed to the normal ASP pattern of lower initial cost and higher annual cost. One
vendor representative said that the high price was partly due to the cost of the Oracle
database license.
The survey data for proprietary ILSs were lower: for this library size category
(10,000 to 25,000 collection items) the mean prerequisite cost was $4,594 («=8), initial
cost was $9,700 («=5), and the annual cost was $3,200 («=10). An exact comparison
with the vendors' quotes is difficult due to the small number of data points and the
differences between the quotes and survey data on vendors, ILS functionality, and
geographic regions. However, given that several of the proprietary ILS respondents had
installed their systems years ago, the price difference might indicate that initial costs of
52
proprietary ILSs are currently higher than reported in this survey.
A Koha consulting firm in the United States estimated the cost for a library of
15,000 items at $10,700 including hardware and the first year of support. Thereafter, the
annual support cost would be $2,500. A lower initial cost for an ASP setup was
estimated at $6,150 with daytime-only support, not including training. Full 24/7 support
and training raised the ASP price to $10,250. The comparable survey data for Koha in
this size category are prerequisite cost of $3,382 (n=17), initial cost of $3,859 (n=16),
and annual cost of $1,429 (n=14).
Comparison of open source and proprietary ILS costs. The cost questions asked
by this survey were prerequisite system requirements (Table B2), initial cost (Table B3),
and annual cost (Table B4). The initial cost was further broken into materials (Table B5)
and labor (Table B6). The most important of the costs is annual cost. The initial costs
are one-time costs, and the prerequisite costs are influenced but not completely
determined by the ILS. Respondents were asked to give answers in terms of cost ranges.
The Tables B2-B6 were computed using the midpoints of these cost ranges, which are
imprecise. The figures are given in U.S. dollars.
The cost questions were not asked of individual Evergreen libraries. The
reported costs for Evergreen libraries were derived by dividing the total consortium costs
by the number of libraries (265) in the Georgia PINES consortium. These figures
represent average costs for individual Evergreen libraries.
ILS cost is affected by library size. The ILS costs of large libraries are more than
those of small libraries. Among those surveyed, the libraries using open source ILSs
53
were on average smaller than those using proprietary ILSs. To control for library size in
Tables B2-B6, the cost data are displayed by library size ranges. The costs of ILSs are
compared within like-sized libraries. Because of the breakdown into smaller subsamples,
most of the differences within the size categories are statistically insignificant.
The figures for initial materials cost (Table B5) are imprecise because no option
for zero cost was offered in the questionnaire. However, the initial materials cost of
most open source ILSs is zero. The lowest cost range in the questionnaire was $0 to
$500 USD. Therefore, the initial materials costs of Koha might actually be as much as
$250 lower than displayed in Table B5.
Except for initial labor cost (Table B6), the average costs of open source ILSs are
less than proprietary ILSs in almost all cost and size categories. Three data points of
low-cost proprietary ILSs for small collections fell outside this pattern and cost less than
their open source counterparts. The higher initial labor costs of open source ILSs in
some categories corroborate the view that OSS requires greater technical sophistication
to use than proprietary software.
Let us consider alternative explanations for the cost findings. Koha libraries in
this survey are found in greater numbers in lower income countries than are libraries
using proprietary ILSs (29% vs. 11%). The lower labor rates in rural areas and
developing countries might cause the Koha costs to appear lower. To control, albeit
crudely, for diverse labor rates of geographic regions, the cost data were filtered on
whether the gross national income per capita of the respondent's country was over or
under $25,000 USD.
54
With this filter, the results were somewhat less clear. Comparing within like
categories of library size and regional income, some proprietary ILSs cost on average
less than Koha on system prerequisites and initial costs, although the difference was
statistically insignificant. This regional income filter did not change the outcome of
annual cost and initial labor cost comparisons: the annual costs of Koha are lower than
those of proprietary ILSs, but the initial labor costs of Koha are sometimes higher.
The costs reported in this survey might be inaccurate for several reasons. The
wording of the question on initial costs was not specific: respondents might or might not
have included costs of migration and customization. Currency conversion to U.S.
dollars may have been inaccurate partly due to fluctuating exchange rates. Also, the
question text on labor costs did not mention internal costs such as staff salaries. Thus
respondents may have omitted internal costs, and this omission might have
disproportionately underreported Koha costs, because more open source ILS libraries
provide ILS services internally. One respondent said, "Koha is the most cost-effective
as far as outlay of money, but the time that it required of me does have a cost that isn't
always factored in. On the other hand, ... my new knowledge has a benefit for the
school district that I work for."
Other characteristics of the sample groups may affect the costs. For example, the
cost per individual Evergreen library is low in part due to the economy of scale achieved
by the consortium. Also, libraries with open source ILSs might be using more volunteer
time to work on them than libraries with proprietary ILSs, and therefore might
underreport the costs of open source ILSs. On the other hand, more libraries using open
55
source customized their ILS, which may have raised their costs. The libraries using
proprietary ILSs may have reported lower costs because their ILSs were purchased
several years ago (n=18), or because of their membership in consortia (n-\3).
A proprietary ILS respondent suggested one of the reasons for the high costs of
proprietary ILSs is the use of proprietary databases: "Commercial ILS vendors have
also become 'married' to relational database vendors such as Oracle to provide back-end
solutions to their products. This will result in a continued escalation in costs and will
bring into question sustainability for many libraries."
Summary of the Findings
The survey asked libraries about their ILS: the reasons for choosing it,
satisfaction, problems, and costs. The main reasons given by libraries for choosing
Koha, Evergreen, and proprietary ILSs were affordability, scalability, and functionality,
respectively. Libraries using open source ILSs more often chose their ILS on the basis
of cost-effectiveness than did libraries using proprietary ILSs. Of libraries that
considered Koha but did not choose it, two thirds found it too complex or difficult to
install, and others found it lacked functionality.
The survey respondents rated their satisfaction with various aspects of their ILS.
Libraries using open source ILSs reported slightly higher overall satisfaction than did
libraries using proprietary ILSs. Koha most exceeded proprietary ILSs in affordability,
but proprietary ILSs most surpassed Koha on installation smoothness. Users of both
Koha and Evergreen were significantly more satisfied than users of proprietary ILSs
with the customizability of their system, but far less satisfied with its documentation
56
completeness.
Besides installation and documentation issues, Koha and proprietary ILSs did not
differ significantly on other problems. However, in its first 15 months of use, Evergreen
had more problems, likely due to its newness, whereas the proprietary ILSs with which it
was compared were older, some by more than 10 years. In this regard it was premature
to survey the users of Evergreen.
Taking library size into account, open source ILSs cost less on average than
proprietary ILSs, except for some initial labor costs. A higher proportion of Koha
libraries were found in countries with lower labor rates. Dividing the data into high and
low income subsamples by national per capita income, some cases were found in which
proprietary ILSs cost less than Koha. Due to small subsample sizes, the differences
were statistically insignificant. The annual cost of Koha was still consistently lower than
proprietary ILSs, taking library size and regional labor rates into account.
The consortium of libraries using Evergreen saved $100,000 annually over its
previous ILS. A great deal of this savings was because OSS can be used with
commodity hardware. The ILS costs were low compared with the other libraries in this
survey also for a reason unrelated to open source: the economy of scale reaped by 265
cooperating libraries.
57
Conclusion
Alternative Explanations and Limitations of the Study
This study sought to understand differences between libraries using open source
and proprietary ILSs. However, because the two sample groups were not otherwise
identical, differences in the results might not be solely due to whether or not the ILS was
open source, but due to differences in other sample characteristics such as library size,
type, location, ILS functionality, whether the ILS was customized, and who provided
services for it. Also, the sample selection procedures for the two sample groups were
different. All participants were recruited by survey invitations posted to listservs, but
open source participants additionally received an individual email.
Some of the differences in the characteristics of the libraries were as follows.
The libraries using open source ILSs in this survey were smaller on average than
libraries using proprietary ILSs. The open source libraries more often hosted their ILS
and provided services internally. The libraries using Evergreen were all public libraries,
part of a consortium in Georgia. By contrast, the libraries using Koha were the most
geographically diverse, located internationally. In this sample, only 12% of libraries
using Koha were public libraries, compared with 30% of libraries using proprietary ILSs.
These underlying differences in the sample groups may compromise the validity
of the results. For example, the higher proportion of services provided internally by
library staff for open source ILSs might raise the satisfaction ratings of open source ILSs
because of library staff members' personal investment in the ILS. To control for some of
these variables, the cost and satisfaction tables were filtered by whether the ILS was
58
customized, where it was hosted, and who provided ILS services. However, no patterns
or logical connections emerged. The subsample sizes were often too small to be
statistically significant for these tests.
Respondents who felt open source ILSs were a more ethical choice might have
rated them higher for ideological reasons rather than practical ones. Also, in the
satisfaction ratings, open source ILS respondents' expectations may have been tempered
by paying lower costs.
The study also had other limitations, which must be taken into account. It did not
compare the ILSs on functionality, other than to ask whether the ILS had modules such
as circulation or serials. Survey respondents reported on their library's ILS and did not
examine other ILSs in direct comparison as respondents in Wilson's (2005) study did.
The respondents in the present survey were mainly library employees; library patrons'
opinions of the usability of the ILS were not included as they were in the GPLS (2007)
survey. Benefits of the ILS were expressed as satisfaction ratings, which are subjective.
The wording of the questionnaire may have not been detailed enough to yield accurate
answers, especially in the questions on costs.
Recommendations for Future Research
The length of the survey was a tradeoff between complete coverage of an ILS
and attention span of a respondent. Splitting the survey into shorter questionnaires
would allow more detailed questions and greater accuracy in answers, while keeping the
length manageable.
59
To give a clearer picture of an ILS, these questions could be added:
• When was the ILS first installed? Last upgraded? • What is the current version number of the ILS? • Is your ILS the one you originally chose, or has your vendor substituted a
different one due to obsolescence or vendor buyout?
A future survey could also include questions to determine if progress has been made on
the weaknesses of open source ILSs found in the present survey, such as installation,
documentation, and the Koha Z39.50 module.
The validity of the survey could be increased by eliminating extraneous factors
where possible. For instance, narrowing the time span of the questions to the current
version of the software would make the comparisons between open source and
proprietary ILSs more equal. Excluding the migration from questions about installation
time and cost would increase the focus on the current ILS, because migration depends on
the previous ILS and the cleanliness of the ILS data.
Improvements to the cost questions could include asking for customization costs
separately and reminding respondents to include internal staff costs for ILS-related
services. Clearer definitions of library collection size and population size may yield
more accurate data. Conducting the survey by telephone rather than on the Web might
increase the accuracy of responses, but would be more expensive.
As more libraries adopt Evergreen, it would be interesting to survey them. Apart
from this survey format, it would also be interesting to investigate the costs of other
large consortia with shared catalogs, in order to compare them with the Georgia PINES
consortium. Comparing Koha with Evergreen was not a goal of this study. However, in
60
the future, if Koha could scale up and Evergreen could scale down, so that enough
similarly sized libraries using them could be surveyed, a comparison might be useful.
Implications for Future Practice
The survey results suggest that open source ILSs need smoother installation
processes and complete documentation. Many respondents encountered parts of the
open source systems that did not work. These issues are significant barriers to adoption
of open source ILSs. Improving these aspects could lower the barrier to their use. Lack
of documentation creates duplication of effort as users encounter the same issues and
independently reinvent solutions. There is no inherent reason why an ILS should be
more difficult to install than an operating system. Open source projects with larger user
and developer bases have been shown to produce higher quality software that is more
functional and beneficial to their communities. Therefore, remedying these issues will
benefit the library community by expanding the user population, who can in turn
contribute feedback, testing, documentation, bug fixes, and code to the ILS.
In the future, it should be possible to create a subset of Evergreen for use in small
libraries. Also, an automated installer that combines the components of Koha was
suggested by several respondents. This would help to eliminate mismatched
components and reduce human error during installation. For small libraries, an ILS that
can coexist with other software on the same machine would save the cost of having to
buy another computer. It is reasonable to hope that installing an open source ILS would
be as smooth and bug-free as installing the LAMP components, which can be done by a
system administrator without debugging. The library community should not expect less.
61
One respondent suggested, "Koha should be guided by an international forum
combining librarians, library school teachers and computer professionals." The open
source ILSs Koha and Evergreen were begun with public funds at public libraries, and
continued development is sponsored by public and nonprofit libraries and universities.
Volunteers also contribute to the projects. However, commercial firms currently control
the development of these open source ILSs.
This arrangement might not be in the best interest of the library community
because the priorities of for profit companies are often to pursue the highest-paying
clients. Their priorities may be different than those of the library and open source
community, whose goals are sharing information and technology. The difficulties of
installation and lack of documentation may profit commercial open source firms by
perpetuating dependence on them.
Therefore, this study recommends the creation of a nonprofit foundation for open
source ILSs. Such a foundation could promote the principles of libraries and open
source, prioritize development based on users' needs, and lower the technical barriers to
use for small, independent libraries. Several such foundations have been set up for other
open source projects (e.g., Apache, Linux, Perl) to coordinate resources and advocate for
their communities.
Discussion
Libraries chose the open source ILS Koha primarily because it was affordable,
and secondarily because it was functional, customizable, and free from vendor lock-in.
Several respondents chose it because it was open source. Some of this group felt it was
62
an ethical choice to "share with other institutions" and "promote cooperative work."
Strengths and weaknesses of OSS commonly cited in the library and open source
literature are confirmed by the survey. Respondents found open source ILSs more
affordable and customizable than proprietary ILSs, but neither as easy to install nor as
well documented. The total cost of an open source ILS, taking into account both initial
and perpetual costs, was on average less than that of a proprietary ILS. However, the
initial cost of labor to install an open source ILS was often higher than that of a
proprietary ILS. The word^ree in free software thus means having liberty to view the
source code, rather than having no cost.
Vendors of proprietary software have criticized OSS, but this criticism is
challenged by the surprising finding that libraries using open source ILSs are actually a
bit more satisfied overall than libraries using proprietary ILSs, although they pay less for
them. The hypothesis of this study is that open source ILSs are more cost-effective than
proprietary ILSs. This is true for libraries that have been able to develop or debug an
open source ILS, or engage a consultant to do so. Their satisfaction is slightly higher
and their costs are lower. However, considering the limitations of this study—that
effectiveness was measured as user satisfaction rather than functionality, and that the
wording of cost questions may not have given rise to accurate answers—further research
is needed.
Libraries that have technical staff persons but few additional funds for computing
stand most to benefit from an open source ILS. How much technical expertise is needed
by libraries in order to use an open source ILS? The goal might simply be one of system
63
administration, which involves installing and maintaining hardware and software. This
requires following detailed instructions, but not programming logic. If libraries take
responsibility and initiative to provide basic technical skills, open source ILS developers
could meet them halfway with software that installs smoothly without requiring
programming and debugging to use.
At the time of this survey, open source ILSs are still works in progress. For
example, the Z39.50 copy cataloging module of the current stable version (2.2.9) of
Koha is not working without debugging, as several respondents noted. This is an
impediment even for experienced system administrators. The incomplete Koha
documentation in the online manual does not cover basics—for example, how to delete a
catalog record. As for Evergreen, which runs on clusters of servers, reports of installing
it in small environments or on a single server are few (LaJeunesse, 2008; Scott, 2008).
Theoretically, there is no reason why an ILS such as Koha should be more difficult to
install than an operating system, nor is there a reason why Evergreen cannot be scaled
down to a small library. These systems might reach their theoretical promise in the
future. However, their development might not be rapid, partly because it is currently
controlled by commercial open source companies whose focus is revenue, rather than the
library community.
The open source library community might be better served by electing a project
leader and release manager from a user group, rather than from a commercial firm whose
interests might conflict with the community. Librarians might form a nonprofit
foundation to keep the development of open source ILSs focused on public service goals.
64
The tasks would be maintaining the scalability and advanced features appropriate to
large libraries and consortia, without neglecting the basics of quality, documentation,
and ease of installation needed by all libraries, including small, independent ones. This
might provide the development of open source ILSs with a democratic balance.
Summary
An open source ILS is a potentially functional, affordable tool. It is particularly
appropriate for libraries because librarians and OSS developers share values of freedom
of information and equal access, and serve their communities by providing materials
economically. This survey found that libraries using open source ILSs choose them
mainly for affordability, and that open source ILSs cost less than proprietary ones.
Users of open source ILS are modestly more satisfied than users of proprietary ILSs.
However, according to this survey, the Koha ILS in its present form lags furthest behind
proprietary ILSs in installation smoothness. Also, the two major open source ILSs,
Koha and Evergreen, compare unfavorably with proprietary ILSs in documentation
completeness.
The early architect of free software, Richard Stallman, compared free software to
public radio, which benefits all users regardless of their ability to pay (2002, p. 59). Has
the potential of an open source ILS been reached? Not yet, for technical and, possibly,
organizational reasons. The technical issues of installation, documentation, and software
quality are yet to be resolved. A recommendation of this study is to focus future
development efforts on alleviating these defects, in order to lower technical barriers to
the use of open source ILSs. Organizationally, open source ILS projects might be best
65
coordinated by a nonprofit entity. Improving these technical and organizational aspects
of open source ILSs might bring open source software closer to its full potential for
libraries.
66
References
American Library Association. (2008, January 22). Code of ethics of the American Library Association. Retrieved April 18, 2008, from http://www.ala.org/ala/oif/statementspols/codeofethics/codeethics.cfm
Antelman, K. (2004, September). Do open-access articles have a greater research impact? College and Research Libraries, 65(5), 372-382. Retrieved August 4, 2007, from http://www.ala.org/ala/acrl/acrlpubs/crliournal/crl2004/crlseptember/antelman.pd f
Apodaca, R. (2006, December 29). Dispelling open source confusion: An introduction to licenses. Retrieved April 1, 2007, from http://depth-first.com/articles/2006/12/29/dispelling-open-source-confusion-an-introduction-to-licenses
ArXiv. (2008a). ArXiv. Retrieved May 17, 2008, from http://www.arxiv.org
ArXiv. (2008b). Mirror sites and other servers: ArXiv mirror sites. Retrieved May 17, 2008, from http://arxiv.org/servers.html
ArXiv. (2008c). ArXiv monthly submission rate statistics. Retrieved April 20, 2008, from http://arxiv.org/show_monthly_submissions
Association of Research Libraries. (2008). Summary statistics. Retrieved April 20, 2008, from http://fisher.lib.virginia.edu/cgi-local/arlbin/arl.cgi?task=setupstats
Balas, J. L. (2005, May). There's no need to fear open source. Computers in Libraries, 25(5), 36-38. Retrieved May 16, 2008, from http://www.infotodav.com/cilmag/may05/balas.shtml
Barbier, B. (2005, August 31). Analyse technique des SIGB Koha et PMB [Technical analysis of the Koha and PMB ILSs]. Unpublished manuscript, Universite de Liege, France.
Bergstra, J. A. & Klint, P. (2005, November 25). About "trivial" software patents: The IsNot case. Retrieved April 19, 2008, from http://homepages.cwi.nl/~paulk/patents/isnot.pdf
Bergstrom, T. C. (2001, March 20). Free labor for costly journals? Retrieved June 9, 2007, from http://www.econ.ucsb.edu/~tedb/Journals/ieprevised.pdf
67
Berizzi, L. & Zweifel, C. (2005). Le pingouin bibliothecaire: Les logiciels libres de gestion de bibliotheque. [The penguin librarian: Free software for library management.] Revue electronique suisse de science de I'information, 2. Retrieved May 18, 2008, from http://campus.hesge.ch/ressi/Numero 2 iuillet2005/articles/PDF/RESSI 010 L B CZ_Pingouin.pdf
Bessen, J. & Hunt, R. M. (2004, March). An empirical look at software patents. Retrieved April 1, 2007, from http://www.researchoninnovation.org/swpat.pdf
Blake, S. (2000, August 24). Some questions from a newbie. Message posted to Koha electronic mailing list, archived at http://lists.katipo.co.nz/public/koha/2000/004402.html
Boss, R. W. (2005, October 19). Open source integrated library system software. Retrieved October 29, 2006, from http://www.ala.org/ala/pla/plapubs/technotes/OpenSource.doc
Breeding, M. (2002, October). An update on open source ILS. Information Today, 19(9), 42-A3. Retrieved December 4, 2005, from http://www.infotoday.com/IT/oct02/breeding.htm
Breeding, M. (2007a). Lib-web-cats. Retrieved September 9, 2007, from http://www.librarvtechnology.org/libwebcats
Breeding, M. (2007b). Lib-web-cats: Current automation system = Koha. Retrieved March 11, 2007, from http://www.librarvtechnology.org/lwc-processquery.pl?ILS=Koha
Bretthauer, D. (2002, March). Open source software: A history. Information Technology and Libraries, 27(1), 3-10. Retrieved May 15, 2008, from http://www.lita.org/ala/lita/litapublications/ital/2101bretthauer.cfm
British Columbia Public Information Network for Electronic Services [BC PINES]. (2007, June 6). Evergreen talking points. Retrieved July 20, 2007, from http://pines.bclibrary.ca/resources/talking-points
Buchanan, K. & Krasnoff, B. (2005, January/February). Can open source software save school libraries time and money? Knowledge Quest, 33(3), 32-34.
Bureau of Labor Statistics. (2008). Inflation calculator. Retrieved April 20, 2008, from http://data.bls.gov/cgi-bin/cpicalc.pl
68
Cana, M. (2003, July 5). Open source and Ranganathan's five laws of library science. Retrieved April 16, 2008, from http://www.kmentor.com/socio-tech-info/2003/07/open-source-and-ranganathans-f.html
Cargile, C. (2005, Spring). Open source ILS for the non-systems librarian: A reality? PNLA Quarterly, 69(3), 15-16. Retrieved May 16, 2008, from http://www.pnla.org/quarterly/Spring2005/PNLA_Spring_05.pdf
Case, M. M. (2004, June). Information Access Alliance: Challenging anticompetitive behavior in academic publishing. College and Research Libraries News, 65(6). Retrieved May 24, 2008, from http://www.ala.org/ala/acrl/acrlpubs/crlnews/backissues2004/iune04/iaa.cfrn
Cervone, F. (2003, Summer). The open source option. Library Journal Net Connect, 8-12. Retrieved May 16, 2008, from http://www.librarviournal.com/article/CA304084.html
Chalon, P., Alexandre-Joaquim, L., Naget, C. & Becquart, C. (2006, September). Open your mind! Selecting and implementing an integrated library system: The open-source opportunity. Paper presented at the 1 Oth European Conference of Medical and Health Libraries, Cluj-Napoca, Romania. Retrieved May 16, 2008, fromhttp://www.eahilconfclui.ro/docs/lb/chalon&al2006 text.pdf
Chudnov, D. (1999, August). Open source software: The future of library systems? Library Journal, 124(13), 40-43.
Clarke, D. (2000, May). Open source software and the library community. Master's thesis, University of North Carolina at Chapel Hill. Retrieved May 16, 2008, from http://www.ils.unc.edu/MSpapers/2576.pdf
Corrado, E. M. (2004, November 9). Comparison of selected open source software library management systems. Retrieved May 16, 2008, from http://www.tcni.edu/~corrado/scholarly/il2004/ossfeatures.html
Corrado, E. M. (2005, Spring). The importance of open access, open source, and open standards for libraries. Issues in Science and Technology Librarianship, 42. Retrieved April 25, 2008, from http://www.istl.org/05-spring/article2.html
Coyle, K. (2002, March). Open source, open standards. Information technologies and libraries, 27(1), 33-36. Retrieved May 16, 2008, from http://www.lita.Org/ala/lita/litapublications/ital/2101 covle.cfm
69
Dingley, B. (2004). U.S. periodical prices - 2004. Retrieved May 16, 2008, from http://ala.org/ala/alctscontent/pubsbucket/alctsresources/general/periodicalsindex /04USPPI.pdf
Do, C. (2007, June 23). Seattle Conference on Scalability: YouTube scalability. Retrieved April 6, 2008, from http://video.google.com/videoplay?docid=-6304964351441328559
Dongarra, J., Golub, G., Grosse, E., Moler, C , & Moore, K. (2006, April). Twenty-plus years of Netlib and NA-Net, Part I. SIAMNews, 39(3), pp. 1, 4. Retrieved May 16, 2008, from http://siam.tekdevelopment.com/old-issues/2006/april-2006/twentv-plus-vears-of-netlib-and-na-net-part-i
European Commission. (2006, January). Study on the economic and technical evolution of the scientific publication markets of Europe. Retrieved April 25, 2008, from http://ec.europa.eu/research/science-society/pdf/scientific-publication-study_en.pdf
Eysenbach, G. (2006). Citation advantage of open access articles. Public Library of Science Biology, 4(5), el 57. Retrieved August 4, 2007, from http://dx.doi.org/10.1371/iournal.pbio.0040157
Federal Research Public Access Act of 2006, S. 2695,109th Cong. (2006, May 2). Retrieved April 25, 2008, from http://thomas.loc.gov/cgi-bin/querv/z?cl09:S.2695:
Federal Trade Commission. (2003, October 28). FTC issues report on how to promote innovation through balancing competition with patent law and policy. Retrieved April 19, 2008, from http://www.ftc.gov/opa/2003/10/cpreport.shtm
Feist Publications, Inc. v. Rural Telephone Service Co., 499 U.S. 340 (1991). Retrieved April 18, 2008, from http://www.law.cornell.edu/copvright/cases/499_US_340.htm
Free Software Foundation. (1991, June). GNU general public license. Retrieved July 11, 2006, from http://www.gnu.org/licenses/gpl.html
Gadd, E., Oppenheim, C. & Probers, S. (2003). RoMEO studies 1: The impact of copyright ownership on academic author self-archiving. Retrieved April 20, 2008, from http://www.lboro.ac.uk/departments/ls/disresearch/romeo/RoMEO%20Studies% 201.pdf
70
Georgia Public Library Service [GPLS]. (2006a). Advantages of Evergreen for PINES library users. Retrieved November 4, 2006, from http://www.georgialibraries.org/lib/pines/Evergreen_Users.doc
GPLS. (2006b). Frequently asked questions. Retrieved November 4, 2006, from http://open-ils.org/faq.php
GPLS. (2007). PINES annual patron satisfaction survey 2007. Retrieved April 6, 2008, from http://www.georgialibraries.org/lib/pines/PINES_Survey_07.pdf
Gorman, M. (1995, September). Five new laws of librarianship. American Libraries, 26(8), 784-785.
Gruener,W. (2007, January 11). IBM leads 2006patent list. Retrieved April 19, 2008, from http://www.tgdaily.eom/content/view/30724/l 18
Hane, P. J. (2003, November 17). Cornell and other university libraries to cancel Elsevier titles. Retrieved April 20, 2008, from http://newsbreaks.infotoday.com/nbreader.asp? ArticleID=16580
Harris Poll. (2006, May 31). Large majorities of U.S. adults support easy-and free-online access to federally-funded research findings on health issues and other topics. The Harris Poll, 44. Retrieved April 25, 2008, from http://www.harrisinteractive.com/harris_poll/index. asp?PID=671
Hedges, S. (2003, October 20). Nelsonvillepublic library: Questions and answers about open source. Retrieved May 16, 2008, from http://www.webiunction.org/do/DisplayContent?id:=l 172
Hedges, S. (2005). A Koha diary: Implementing Koha at the Nelsonville public library. Retrieved May 16, 2008, from http://www.kohadocs.org/koha_diary.html
Hoff, T. (2008, March 13). YouTube architecture. Retrieved April 6, 2008, from http://highscalability.com/youtube-architecture
House of Commons Science and Technology Committee. (2004, 20 July). Scientific publications: Free for all? Retrieved July 22, 2007, from http://www.publications.parliament.uk/pa/cm200304/cmselect/cmsctech/399/399 .pdf
Ibiblio. (n.d.). Ibiblio frequently asked questions for visitors. Retrieved April 8, 2007, fromhttp://www.ibiblio.org/faq/?sid=l
71
Jefferson, T. (1813, August 13). To Isaac McPherson. In P. B. Kurland and R. Lerner (Eds.), The founder's constitution. (1986). Chicago: University of Chicago Press. Retrieved April 18, 2008, from http://press-pubs .uchicago.edu/founders/documents/a 1 8 8 s 12 .html
Jefferson, T. (1818, March 3). To Dr. Benjamin Waterhouse. In P. L. Ford (Ed.), The works of Thomas Jefferson: Vol. 12. (Federal ed., 1905, p. 88). New York: G. P. Putnam's Sons. Retrieved April 18, 2008, from the Online Library of Liberty, http://oll.libertyfund.org/title/808/88371
Kirby, R. (1997, May 27). Math journal prices. Retrieved June 9, 2007, from http ://math.berkeley. edu/~kirby/j ournals .html
Knuth, D. E. (1994). Letter to the Patent Office from Professor Donald Knuth. Retrieved April 19, 2008, from http://progfree.org/Patents/knuth-to-pto.txt
Koha Developer Wiki. (2007). Koha users around the world. Retrieved November 4, 2006, from http://wiki.koha.org/doku.php?id=kohausers
KSR International Co. v. Teleflex Inc. et al., 127 S. Ct. 1727 (2007). Retrieved April 19, 2008, from http://www.supremecourtus.gov/opinions/06pdf/04-1350.pdf
Kumar, V. (2005). Free/open source integrated library management systems: Comparative analysis of Koha, PhpMyLibrary, and OpenBiblio. Paper presented at the National Library Week Celebration, University of Calicut. Retrieved May 16, 2008, from http://eprints.rclis.org/archive/00008074
Kutz, M. (2002, January). The scholars' rebellion against scholarly publishing practices: Varmus, Vitek, and venting. Searcher, 10(1). Retrieved April 20, 2008, from http://www.infotoday.com/searcher/ian02/kutz.htm
LaJeunesse, B. (2005, April). Evergreen: Built for a consortium. Library Journal Net Connect. Retrieved May 16, 2008, from http://www.libraryiournal.com/article/CA515806.html
LaJeunesse, B. (2006). Evergreen-user: Installing and running the staff client. Retrieved Nov. 4, 2006, from http://www.open-ils.org/dokuwiki/doku.php?id=evergreen-user:installing_and_running_the_staff_client
LaJeunesse, B. (2008, February 11). Isn 't Evergreen just for consortia with hundreds of libraries? Retrieved May 6, 2008, from http://open-ils.org/blog/?p=120
72
LibLime. (2005). Liblime partners with Index Data to enhance Koha. Retrieved May 16, 2008, from http://liblime.com/news-items/press-releases/liblime-partners-with-index-data-to-enhance-koha
Linux Counter. (2005, March). Estimating the number of Linux users. Retrieved April 12, 2008, from http://counter.li.org/estimates.php
McVeigh, M. E. (2004). Open access journals in the ISI citation databases: Analysis of impact factors and citation patterns. Retrieved April 21, 2008, from http://scientific.thomson.eom/media/presentrep/essayspdf/openaccesscitations2.p df
Megginson, D. (2000). Copyright status. Retrieved April 1, 2007, from http://www.saxproject.org/copying.html
Michigan Evergreen. (2007, October 18). Welcome to the Michigan Evergreen project. Retrieved April 6, 2008, from http://www.mlcnet.org/evergreen/?p=3
Moczar, L. (2005). The economics of commercial open source. Retrieved April 6, 2008, from http://www.galatea.com/economics.html
Murray, F. & Stern, S. (2005, June). Do formal intellectual property rights hinder the freeflow of scientific knowledge? An empirical test of the anti-commons hypothesis. Retrieved April 19, 2008, from http://imio.haas.berkeley.edu/WilliamsonSeminar/sternl 11705.pdf
National Institutes of Health. (2008). National Institutes of Health public access. Retrieved April 18, 2008, from http://publicaccess.nih.gov
National Science Foundation Office of Inspector General. (2006, February 17). Audit of NSF's policies on public access to the results of NSF-funded research (OIG 06-2-004). Retrieved June 10, 2007, from http://www.nsf.gov/oig/06-2-004 final.pdf
Netcraft. (2007, August). August 2007'Web server survey. Retrieved April 12, 2008, from http://news.netcraft.eom/archives/2007/08/06/august_2007_web_server_survey.h tml
Netcraft. (2008, February). Tiscali Italia is the most reliable hosting company in February 2008. Retrieved April 12, 2008, from http://news.netcraft.com/archives/2008/03/04/tiscali_italia_is_the_most_reliable_ hosting_company_in_february_2008.html
73
Netcraft. (2008, March). March 2008 Web server survey. Retrieved April 12, 2008, from http://news.netcraftxom/archives/2008/03/26/march_2008_web_server_survey.h tml
Open Source Initiative. (2005). History of the OSI. Retrieved December 14, 2005, from http://opensource.org/history
Open Source Initiative. (2006). The open source definition. Retrieved October 1, 2006, from http://www.opensource.org/docs/definition.php
Pace, A. K. (2005, September). Helping you buy: Integrated library systems. Computers in Libraries, 25(8), 25-32. Available from http://connection.ebscohost.com/content/article/1039964484.html
Proffitt, B. (2002, August 30). Koha: A library checks out open source. Retrieved December 14, 2005, from http://www.linuxplanet.eom/linuxplanet/reports/4408/l
Public Access to Science Act, HR. 2613, 108th Cong. (2003, June 26). Retrieved April 25, 2008, from http://thomas.loc.gOv/cgi-bin/query/z7cl08:H.R.2613:
Ranganathan, S. R. (1931). The five laws of library science. Madras, India: Madras Library Association. Retrieved April 18, 2008, from http://dlist.sir.arizona.edu/1220
Raymond, E. S. (2001). The cathedral and the bazaar. Sebastopol, CA: O'Reilly.
Red Hat, Inc. (2005a). Does Red Hat distribute under a license other than the GPL? Retrieved April 20, 2008, from http://kbase.redhat.com/faq/FAQ 61 5499.shtm
Red Hat, Inc. (2005b). What is the GPL? Retrieved April 20, 2008, from http://kbase.redhat.com/faq/FAQ 61 5496.shtm
Sampat, B. (2006, July). Patenting and U.S. academic research in the 20th century: The world before and after Bayh-Dole. Research Policy, 35(6), 772-789. Available from http://dx.doi.Org/10.1016/i.respol.2006.04.009
Schwarz, G. & Kennicutt, R. (2004, November 10). Demographic and citation trends in journal papers and preprints. Retrieved July 13, 2007, from http://arxiv.org/abs/astro-ph/0411275
74
Scott, D. (2008, April 14). Tuning PostgresSQL for Evergreen on a test server. Retrieved May 6, 2008, from http://www.co ffeecode.net/archives/2008/04/C 17.html
Seventh Research Framework Programme. (2007). FP7 grant agreement: Annex II, general conditions. Retrieved April 25, 2008, from ftp://ftp.cordis.europa.eu/pub/fp7/docs/fp7-ga-annex2_en.pdf
Snoddy, R. (1997, March 13). Reed Elsevier forward 10%. Financial Times, p. 24.
Stallman, R. (1999). The GNU operating system and the free software movement. In C. DiBona, S. Ockman, & M. Stone (Eds.), Open sources: Voices from the open source revolution (pp. 53-70). Sebastopol, CA: O'Reilly. Retrieved October 29, 2006, from http://www.oreillv.com/catalog/opensources/book/stallman.html
Stallman, R. (2002). No sir, no monopoly. Hyderabad, India: Prajasakti Book House.
Suber, P. (2006, November 12). Journal declarations of independence. Retrieved May 26, 2008, from http://www.earlham.edU/~peters/fos/lists.htm#declarations
Sun Microsystems. (2004). Package org.xml.sax. Retrieved April 1, 2007, from http://j ava. sun, com/j 2 se/1.5.0/docs/api/org/xml/sax/package-summary.html
Swan, A. (2005). Open access self-archiving: An introduction. Retrieved April 20, 2008, from http://www.keyperspectives.co.uk/openaccessarchive/reports/Open%20access%2 0self-archiving%20-%20an%20introduction.pdf
Tennant, R. (2000a, January 1). The role of open source software. Library Journal, 725(1), 36. Retrieved May 16, 2008, from http://www.libraryiournal.com/article/CAl 56513.html
Tennant, R. (2000b, April 15). Technology decision-making: A guide for the perplexed. Library Journal, 125(1), 30. Retrieved May 16, 2008, from http://www.librarviournal.com/article/CAl 56516.html
Tennant, R. (2003, August 15). Open source goes mainstream. Library Journal, 725(13), 30. Retrieved May 16, 2008, from http://www.libraryiournaI.com/article/CA315185.html
Torvalds, L. & Diamond, D. (2001). Just for fun: The story of an accidental revolutionary. New York: HarperBusiness.
75
TuneQuest. (2007, February 26). A big list of MP 3 patents. Retrieved May 17, 2008, from http://www.tunequest.org/a-big-list-of-mp3-patents/20070226
U.S. Const., art. I, § 8, cl. 8. Retrieved May 17, 2008, from http://www.usconstitution.net/const.html#AlSec8
Walter, P. & Yamamoto, K. (2003). Call for boycott of Cell Press journals. Message posted to Budapest Open Access Initiative mailing list, archived at http://threader.ecs.soton.ac.uk/lists/boaiforum/168.html
Weir, R. (2005). Koha on Windows. Retrieved December 14, 2005, from http://www.koha.rwir.com/Koha_on_Windows.html
Wellcome Trust. (2008, February). Position statement in support of open and unrestricted access to published research. Retrieved April 25, 2008, from http://www.wellcome.ac.uk/About-us/Policy/Spotlight-issues/Open-access/Policy/index.htm
Williams, S. (2002). Free as in freedom: Richard Stallman's crusade for free software. Sebastopol, CA: O'Reilly. Retrieved June 14, 2008, from http://oreilly.com/openbook/freedom
Willinsky, J. (2005, August). The unacknowledged convergence of open source, open access, and open science. First Monday, 10(2). Retrieved July 20, 2007, from http ://www. firstmonday. or g/Issues/issue 10_8/willinsky
Wilson, A. (2005, October 14). Review of library management systems. Retrieved May 17, 2008, from http://www.eclaksa.nhs.uk/Uploads/26/2005 10 14%20Review%20of%20LMS. doc
World Intellectual Property Organization. (1979). Berne Convention for the protection of literary and artistic works. Retrieved April 1, 2007, from http ://www.wipo.int/treaties/en/ip/berne/trtdocs_woOO 1 .html
76
Appendix A: Survey Questionnaire1
Introduction
1. Which ILS does your library use?
2. How would you rate your satisfaction with your ILS overall?
1 - very 2 - somewhat 3 - neither 4 - somewhat 5 - very don't know/ dissatisfied dissatisfied satisfied nor satisfied satisfied no answer
dissatisfied
3*. Before your ILS was selected, did the selector(s) analyze costs and benefits of ILS alternatives?
• Yes, this ILS was selected as most cost-effective. • No. • Don't know/no answer • Yes, but this ILS was selected for other reasons (What reasons? Please specify.)
4*. Please rank in order the top 5 reasons your library chose this ILS. Choose "1" for most important, "2" for second most important, etc.
security combined with other incentives desirable features/functionality trusted brand name affordable price good technical support portability ease of use customizability we used this ILS or vendor before freedom from vendor lock-in scalability other (please specify and rate as 1 - 5) don't know/no answer
1 Questions marked with an asterisk (*) were not asked of individual Evergreen libraries, but were asked of the Evergreen consortium director, and all other respondents.
77
5. How would you rate your satisfaction with your ILS on the following items?
1 - very 2 - somewhat 3 - neither 4 - somewhat 5 - very don't know/ dissatisfied dissatisfied satisfied nor satisfied satisfied no answer
dissatisfied
• portability • scalability* • ease of use • affordability* • customizability • features/functionality • technical support • security
• interoperability
6*. Is your ILS hosted on site at your institution or off site?
on site/parent site off site/service provider
7*. Who provides the following items for your library's ILS?
internal staff external contractor both
• Installation • Documentation • Training • Support • Maintenance
Customization
8*. Has your library customized (or sponsored customization of) the ILS source code?
yes no don't know/no answer
9*. Were your customizations
A) limited to modifying the appearance of web pages (in HTML or CSS) such as to show your library name, logo, etc, or to configuring ILS options such as branch libraries and circulation policies,
2 Customization questions 9 and 10 were asked as an email follow-up for Koha respondents.
78
don't know/no answer
don't know/no answer
B) or did your customizations involve changing the source code (in Perl, PHP, or another programming language)
C) or something else?
D) don't know/no answer
10*. Can you please briefly describe your customization?
Installation
The process of ILS installation includes configuration of ILS catalog format and circulation policies, and migration or data entry of the catalog and user accounts.
11*. Were you involved in the installation of your ILS, either directly (installing it yourself) or indirectly (having others install it)?
yes no
12*. How smoothly did the installation of your ILS go? (The process of ILS installation includes configuration of ILS catalog format and circulation policies, and migration or data entry of the catalog and user accounts.)
1 - with 2 - with some 3 - neither 4 - somewhat 5 - extremely don't extreme difficulty smoothly nor smoothly smoothly know/no difficulty with difficulty answer
13*. Approximately how long did it take to install and configure the ILS, including catalog and user account migration? Please enter the approximate number of days.
14*. How acceptable to you was the amount of time it took to install the ILS?
1 - very 2 - somewhat 3 - neither 4 - somewhat 5 - very don't unacceptable unacceptable acceptable nor acceptable acceptable know/no
unacceptable answer
15*. Do you have any comments about the ILS installation process or how it could be improved?
79
OPAC
16. The Online Public Access Catalog (OPAC) contains a search engine to the catalog. Are you familiar with your ILS's OPAC?
yes no
17. Regarding the ILS's OPAC search engine, how would you rate its ease of use?
1 - very 2 - somewhat 3 - neither 4 - somewhat 5 - very don't hard to use hard to use easy nor hard easy to use easy to use know/no
answer
18. How would you rate the OPAC search engine's completeness of features?
1 - lacks all 2 - lacks most 3 - has some 4 - has most 5 - has all don't features we features we but lacks features we features we know/no need need some need need answer 19. What is your perception of how easy it is for patrons to make self-service requests online for renewals and holds?
1 - very 2 - somewhat 3 - neither 4 - somewhat 5 - very don't know/
hard hard easy nor hard easy easy no answer
Circulation and Patron Accounts
The following questions are about your daily experience using your library's ILS.
20. The circulation and patron accounts modules may be used to process book loans, returns, renewals, holds, to enter new patron accounts, resolve fines, or edit patron information. Are you familiar with your ILS's circulation and patron accounts modules?
• Yes. • No, I use them on average less than two hours per week. • No, our ILS does not have circulation and patron accounts modules.* • No, our library is noncirculating.*
21. How would you rate the ease of use of the circulation and patron accounts modules?
1 - very 2 - somewhat 3 - neither 4 - somewhat 5 - very don't hard to use hard to use easy nor hard easy to use easy to use know/no
answer
80
22. How would you rate the completeness of features of the circulation and patron accounts modules?
1 - lacks all 2 - lacks most 3 - has some 4 - has most 5 - has all don't features we features we but lacks features we features we know/no need need some need need answer
23. How easy is it for staff to process loans and returns in the ILS?
1 - very 2 - somewhat 3 - neither 4 - somewhat 5 - very don't know/
hard hard easy nor hard easy easy no answer
24. How easy is it for staff to process requests for renewals and holds in the ILS?
1 - very 2 - somewhat 3 - neither 4 - somewhat 5 - very don't know/ hard hard easy nor hard easy easy no answer 25. How easy is it for staff to do tasks such as adding a new patron account, resolving a fine, or updating a patron's address in the ILS?
1 - very 2 - somewhat 3 - neither 4 - somewhat 5 - very don't know/
hard hard easy nor hard easy easy no answer
Cataloging
26. Are you familiar with your ILS's cataloging module?
• Yes. • No, I use it on average less than two hours per week.
• No, our ILS does not have a cataloging module.*
27. How would you rate the ease of use of the cataloging module?
1 - very 2 - somewhat 3 - neither 4 - somewhat 5 - very don't hard to use hard to use easy nor hard easy to use easy to use know/no
answer completeness of features of the cataloging module? 28. How would you rate the
1 - lacks all 2 - lacks most features we features we need need
3 - has some 4 - has most but lacks features we some need
5 - has all don't features we know/no need answer
81
29. How easy is it to add original catalog records from scratch in the ILS?
1 - very 2 - somewhat 3 - neither 4 - somewhat 5 - very don't know/ hard hard easy nor hard easy easy no answer
30. How easy is it to do copy cataloging in the ILS, that is, to do tasks such as retrieving catalog records from OCLC or Library of Congress and adding local holdings information?
1 - very 2 - somewhat 3 - neither 4 - somewhat 5 - very don't know/ hard hard easy nor hard easy easy no answer
Serials
31*. The serials module may be used to manage journal or magazine subscriptions. Are you familiar with your ILS's serials module?
• Yes. • No, I use it on average less than two hours per week. • No, our ILS does not have a serials module.
32*. How would you rate the ease of use of the serials module?
1 - very 2 - somewhat 3 - neither 4 - somewhat 5 - very don't hard to use hard to use easy nor hard easy to use easy to use know/no
answer
33*. How would you rate the completeness of features of the serials module?
1 - lacks all 2 - lacks most 3 - has some 4 - has most 5 - has all don't features we features we but lacks features we features we know/no need need some need need answer
Acquisitions
34*. The acquisitions module may be used to manage book or materials orders. Are you familiar with your ILS's acquisitions module?
• Yes. • No, I use it on average less than two hours per week. • No, our ILS does not have an acquisitions module.
82
35*. How would you rate the ease of use of the acquisitions module?
1 - very 2 - somewhat 3 - neither 4 - somewhat 5 - very don't hard to use hard to use easy nor hard easy to use easy to use know/no
answer
36*. How would you rate the completeness of features of the acquisitions module?
1 - lacks all 2 - lacks most 3 - has some 4 - has most 5 - has all don't features we features we but lacks features we features we know/no need need some need need answer
Documentation
37. ILS documentation may include manuals, documents, or help menus in your ILS. Have you read any of your ILS documentation?
yes no
38. How would you rate your satisfaction with your ILS documentation (which may include manuals, documents, or help menus)?
1 - very 2 - somewhat 3 - neither 4 - somewhat 5 - very don't know/ dissatisfied dissatisfied satisfied nor satisfied satisfied no answer
dissatisfied
39. How would you rate the completeness of your ILS documentation?
1 - very 2 - somewhat 3 - neither 4 - somewhat 5 - very don't know/ incomplete incomplete complete complete complete no answer
nor incomplete
40. Do you have any comments about the documentation or how it could be improved?
83
Support, Maintenance, and Reliability
41. Support refers to the help you need when you are troubleshooting or seeking clarification on how to use a feature that is new to you. How would you rate your satisfaction with your ILS support?
1 - very 2 - somewhat 3 - neither 4 - somewhat 5 - very have not dissatisfied dissatisfied satisfied nor satisfied satisfied received
dissatisfied support
42. Do you have any comments about your ILS support or how the support could be improved?
43*. Maintenance refers to periodic backups, patches, and upgrades. How would you rate your satisfaction with your ILS maintenance?
1 - very 2 - somewhat 3 - neither 4 - somewhat 5 - very don't know/ dissatisfied dissatisfied satisfied nor satisfied satisfied no answer
dissatisfied
44. How would you rate your satisfaction with your ILS in terms of reliable, trouble-free operation?
1 - very 2 - somewhat 3 - neither 4 - somewhat 5 - very don't know/ dissatisfied dissatisfied satisfied nor satisfied satisfied no answer
dissatisfied
ILS Problems
45. Have you had any problems, such as software bugs, with your ILS?
• Yes, we have had some ILS problems. • Yes, but not sure if the problem was due to something else. • No, we have had no ILS problems.
46. What types of ILS problems have you had (not including problems caused by underlying components such as network failure)? Select all that apply.
• Lapse in service • Data loss • Problems with OPAC/search engine • Problems with circulation module • Problems with cataloging module
84
• Problems with serials module* • Problems with acquisitions module* • Security breach • Other (please specify)
Library Size
47*. How many books and serial volumes are in your library's collection? Your best estimate is fine.
48*. What is the population of your library's service area? Your best estimate is fine.
Costs
Can you please estimate the following costs in U.S. dollars?
49*. What was the cost of any new hardware, software, or network equipment required before installing your ILS?
The next two questions break down the ILS costs by initial and perpetual costs.
50*. The initial cost of the ILS may include the software, license fees, installation, configuration, migration of data from previous ILS, documentation, and training. What was the initial cost of your ILS?
51*. The perpetual costs of the ILS may include support contracts, maintenance, and license fees. What is the annual cost of your ILS?
The next two questions further break down the initial ILS cost into materials and labor costs.
52*. The initial materials cost may include software, documentation, and license fees. Can you please estimate the initial materials cost of your ILS?
53*. The initial labor cost of your ILS may include installation, support, and training. Can you please estimate the initial labor cost of your ILS?
The answer ranges for the cost questions, Q49-Q53, are:
• $0 - $500 USD • $501-$1,000 USD • $1,001-$2,500 USD • $2,501 - $5,000 USD
85
• $5,001 - $7,500 USD . $7,501-$10,000 USD • $10,001-$25,000 USD . $25,001 - $50,000 USD • $50,001 - $75,000 USD • $75,001-$100,000 USD • $100,001-$500,000 USD . $500,001-$1,000,000 USD . $1,000,001 - $5,000,000 USD . $5,000,001 - $ 10,000,000 USD . over $10,000,000 USD • included in cost of ILS (Q. 47 only) • don't know/no answer • can give a closer estimate
Demographics
54. What is the name of your library?
55. What type is your library?
public academic special school other (please specify)
56. Where is your library located?
City: State/Province:
Country:
Finish
57. Do you have any comments about your library's ILS or how it could be
improved?
86
Appendix B:
Table Bl
ILS
Sati
sfac
tion
R
atin
gs
Appendix B:
Tables B1-B6
00
^1
Installation smoothness
Documentation completeness
Installation time
Orig cataloging ease of use
Cataloging ease of use
Reliability
Copy cataloging ease of use
Serials completeness
Technical support
Documentation
Circ/pat completeness
Circ/pat loans s returns
Cataloging completeness
Maintenance
Acquisitions completeness
Circ/pat patron updates
Circ/pat renewals s holds
Acquisitions ease of use
Serials ease of use
Security
Ease of use
OPAC ease of use
Overall ILS satisfaction
OPAC self-service
Circ/pat ease of use
Scalability
OPAC completeness
Features/functionality
Portability
Interoperability
Customizability
Affordability
Subt
n
167
114
133
217
123
205
168
196
190
131
194
191
205
228
219
228
197
196
219
227
176
192
219
otal
ave
2.8
3.7
3.9
3.9
4. 1
3.7
3.1
3.9
4.5
4.0
4.3
4.3
4.2
4.0
4.1
4.2
3.9
4.3
3.9
4.1
4.0
4.0
4.0
Open Source
Koha
n 100
82
99
93
97
104
89
38
97
83
86
84
96
93
35
85
82
35
38
99
112
107
113
88
86
86
107
112
89
100
109
106
ave
2.9
2.7
3.7
3.7
3.9
4.0
4.1
3.4
3.5
3.1
4.1
4.7
4.2
3.9
3.7
4.4
4.3
3.7
3.4
4.0
4.0
4.3
4.3
3.9
4.4
4.3
3.9
4.2
4.1
4.0
4.3
4.6
Evrg
n la
85 la
21
36
113
34
-108
85
110
106
35 1"
-109
109
- -106
116
112
115
109
110 1"
112
115
87
92
110 la reen
ave
5.0
3.0
5.0
3.5
3.9
3.8
4.1
-3.8
3.1
3.8
4.3
3.5
5.0
-4.2
4.2
- -4.4
4.0
3.8
4.1
3.8
4.2
5.0
3.8
4.0
3.9
3.9
3.8
4.0
Proprtry
n 77
107
78
90
95
114
90
54
109
109
105
103
97
106
49
105
102
47
52
114
128
118
126
105
104
102
118
128
89
96
125
120
ave
3.7
3.4
4.2
4.1
4.1
4.1
4.2
3.6
3.8
3.2
3.9
4.4
4.0
3.8
3.5
4.2
4.1
3.4
3.0
4.0
3.8
3.9
3.9
3.5
3.9
3.8
3.6
3.7
3.5
3.5
3.4
3.6
Open Source
vs.
Proprietary
d
-0.6
-0.4
-0.2
-0.2
-0.1
-0.1
-0.1
0.0
0.1
0.1
0.1
0.1
0.2
0.2
0.2
0.3
0.4
0.3
0.3
0.4
0.5
0.5
0.6
P
.000
.012
.100
.130
.454
.503
.592
.718
.610
.502
.236
.206
.121
.104
.052
.029
.007
.002
.002
.001
.001
.000
.000
Koha
vs.
Proprietary
d
-0.8
-0.7
-0.5
-0.3
-0.2
-0.1
-0.1
-0.2
-0.2
-0.1
0.2
0.2
0.2
0.1
0.2
0.2
0.2
0.3
0.3
0.0
0.2
0.5
0.4
0.4
0.5
0.4
0.3
0.5
0.7
0.5
0.8
1.1
P
.000
.000
.008
.036
.123
.683
.500
.468
.175
.552
.087
.026
.026
.3 62
.248
.047
.107
.152
.148
.999
.161
.000
.008
.006
.000
.003
.001
.000
.000
.000
.000
.000
Evergreen
Prop:
d
-0.4
-0.6
-0.2
-0.3
-0.1
0.0
-0.1
-0.1
-0.1
-0.4
0.0
0.1
0.4
0.2
-0.0
0.2
0.3
0.3
0.2
0.3
0.4
0.4
0.3
VD .
rietary
P
.013
.052
.330
.034
.603
.873
.740
.385
.429
.009
.721
.476
.007
.170
.958
.231
.035
.038
.031
.034
.034
.011
.026
Not
e.
Table is sorted by significance of difference of Open Source vs. Proprietary.
n
= number of responses;
ave = mean satisfaction rating; d = difference in satisfaction ratings between sample groups; p = two-tailed
probability value based on Welch's t-test for samples with unequal variances.
'Represents a consortium of 265 libraries.
Appendix B: Table B2
Syst
em
Pre
requ
isit
es
Cos
t of
IL
S
00
00
Collection size
100 items
100 and < 500 items
500 and < 1,000 items
1,000 and < 10,000 items
>= 10,000 and < 25,000 items
>= 25,000 and < 50,000 items
>= 50,000 and < 100,000 items
100,000 and < 1,000,000 items
1,000,000 items
< >=
>=
>=
>=
>=
Population size
< 100 patrons
>= 100 and < 500 patrons
>= 500 and < 1,000 patrons
>= 1,000 and < 10,000 patrons
>= 10,000 and < 25,000 patrons
>= 25,000 and < 50,000 patrons
>= 50,000 and < 100,000 patrons
>= 100,000 and < 1,000,000 patrons
>= 1,000,000 patrons
Open Source
Koha
1
5
39
17 5
4
12
$ 2
3
2 4
3
250
050
981
382
150
750
250
11
24
13
21 5
5 1
1
$1 1
4
2
2
2 1
250
083
115
393
350
450
750
250
Evergreen
265
$566
265
$566
Proprietary
Koha
vs.
Proprietary
d p
1
3 8
8
11
18
10
$
250
6,000
4,594
2,375
48,318
43,417
426,275
$ -5
-1
-43
-40
019
211
-225
568
167
475
689
890
133
023
6
7
12 9
10 5 9
1
$ 3
4
33
11
70
632
127
87
333
571
417
611
150
000
028
500
$ -2
-31
-9
-67
250
-456
024
261
700
.089
.905
.231
.117
.037
Evergreen
vs .
Proprietary
-$1,809
.112
-$32,851
.206
Not
e.
n = number of responses; ave = mean cost; d = difference in cost between sample groups;
p
probability based on Welch's t-test for samples with unequal variances.
two-tailed
Appendix B:
Table B3
Init
ial
Cos
t of
IL
S
Collection size
< 100 items
>= 100 and < 500 items
>= 500 and < 1,000 items
>= 1,000 and < 10,000 items
>= 10,000 and < 25,000 items
>= 25,000 and < 50,000 items
>= 50,000 and < 100,000 items
>= 100,000 and < 1,000,000 items
>= 1,000,000 items
Open Source
Koha
1
5
35
16 6
5
10
$ 1,
1,
3,
5,
19,
21, 250
,950
,407
,859
,958
,250
,675
Evergreen
265
S953
Proprietary
4 $ 25,125
5
9,700
7
8,000
11
78,364
16
217,609
8
266,719
Koha
vs.
Proprietary
-23,718
.218
-5,841
.458
-2,042
.727
-59,114
.127
-195,934
.005
Evergreen
vs .
Proprietary
d
p
-$7,047
.210
00
Population size
< 100 patrons
>= 100 and < 500 patrons
>= 500 and < 1,000 patrons
>= 1,000 and < 10,000 patrons
>= 10,000 and < 25,000 patrons
>= 25,000 and < 50,000 patrons
>= 50,000 and < 100,000 patrons
>= 100,000 and < 1,000,000 patrons
>= 1,000,000 patrons
11
24
12
18 4
5 1
1
$ 1,
2,
2,
7,
2,
37,
17,
1, ,386
.677
,875
,472
,500
.700
,500
,750
265
$953
5 $ 5,500
$
-2,823
.423
5
12 8
7
7
7 1
13,
68,
56,
231,
164,
364,
300, ,150
,250
,125
,607
,286
,286
,000
-10,
-60,
-53,
-193, 275
778
625
907
.209
.084
.177
.095
-$67,297
.059
Not
e.
n = number of responses; ave = mean cost; d = difference in cost between sample groups; p = two-tailed
probability based on Welch's t-test for samples with unequal variances.
Appendix B:
Table B4
Ann
ual
Cos
t of
IL
S
Collection size
< 100 items
>= 100 and < 500 items
>= 500 and < 1,000 items
>= 1,000 and < 10,000 items
>= 10,000 and < 25,000 items
>= 25,000 and < 50,000 items
>= 50,000 and < 100,000 items
>= 100,000 and < 1,000,000 items
>= 1,000,000 items
Open Source
Koha
2 5
35
14 6
5
11
$ 1,
1,
2,
6, 250
450
664
,429
,917
,550
,023
Evergreen
265
$1,698
Proprietary
Koha
vs.
Proprietary
d P
Evergreen
vs.
Proprietary
6
10 8
12
25
10
$ 4,708
3,200
10,281
18,167
38,870
69,250
$ -4,044
-1,771
-8,365
-15,617
-32,847
.203
.157
.123
.017
.012
-$8,583
.10
Population size
O
< 100 patrons
>= 100 and < 500 patrons
>= 500 and < 1,000 patrons
>= 1,000 and < 10,000 patrons
>= 10,000 and < 25,000 patrons
>= 25,000 and < 50,000 patrons
>= 50,000 and < 100,000 patrons
>= 100,000 and < 1,000,000 patrons
>= 1,000,000 patrons
10
24
11
19 5
5 1
1
S
2,
9,
17, 300
813
568
,066
550
,200
,500
250
265
$1,698
6
8
17
10
11 7
11 1
$ 2,583
7,156
11,691
15,275
57,205
30,357
65,568
37,500
$ -1,
-6,
-9,
-14,
-48, 771
588
625
725
.005
.259
.184
.031
.006
.094
-$9,993
.024
Not
e.
n = number of responses; ave = mean cost; d = difference in cost between sample groups; p = two-tailed
probability based on Welch's t-test for samples with unequal variances.
Appendix B:
Table B5
Init
ial
Mat
eria
ls
Cos
t of
IL
S
Open Source
Koha
Evergreen
Proprietary
Koha
vs.
Proprietary
Evergreen
vs.
Proprietary
Collection size
< 100 items
>= 100 and < 500 items
>= 500 and < 1,000 items
>= 1,000 and < 10,000 items
>= 10,000 and < 25,000 items
>= 25,000 and < 50,000 items
>= 50,000 and < 100,000 items
>= 100,000 and < 1,000,000 items
>= 1,000,000 items
Population size
v©
< 100 patrons
*""
>= 100 and < 500 patrons
>= 500 and < 1,000 patrons
>= 1,000 and < 10,000 patrons
>= 10,000 and < 25,000 patrons
>= 25,000 and < 50,000 patrons
>= 50,000 and < 100,000 patrons
>= 100,000 and < 1,000,000 patrons
>= 1,000,000 patrons
1
5
36
15 6
3 9
$
250
350
306
383
250
3,083
250
11
23
11
18 4
4
1
1
$
295
337
295
806
250
250
250
250
265
$0
265
$0
2 $
250
56
.211
5 5 6
7 4
4,
3,
40,
44,
97, ,700
,150
,708
,821
,375
-4,
-2,
-37,
-44, 317
900
625
571
.257
.126
.089
.002
2 3
8
5 4
3 3
1
$
1,000
5,917
8,063
14,200
112,563
47,500
79,167
1,750
$
-5,
-7,
-13,
-112, -663
, 621
,257
,950
,313
.540
.186
.168
.313
.183
-$3,150
.104
.063
.130
Not
e.
n = number of responses; ave = mean cost;
d
= difference in cost between sample groups; p = two-tailed
probability based on Welch's t-test for samples with unequal variances.
Appendix B:
Table B6
Init
ial
Lab
or
Cos
t of
IL
S
Collection size
< 100 items
>= 100 and < 500 items
>= 500 and < 1,000 items
>= 1,000 and < 10,000 items
>= 10,000 and < 25,000 items
>= 25,000 and < 50,000 items
>= 50,000 and < 100,000 items
>= 100,000 and < 1,000,000 items
>= 1,000,000 items
Open Source
Koha
2 5
37
14 6
3
11
$
250
1, 950
1,358
3,232
5,958
9,167
18,023
Evergreen
265
$953
Proprietary
Koha
vs.
Proprietary
2 5
4
5
8
5
$
250
1,350
375
14,600
34,719
139,300
$
1,108
1,882
5,583
-5,433
-16,696
.002
.205
.084
.532
.218
Evergreen
vs.
Proprietary
d p
578
.019
Population size
< 100 patrons
>= 100 and < 500 patrons
>= 500 and < 1,000 patrons
>= 1,000 and < 10,000 patrons
>= 10,000 and < 25,000 patrons
>= 25,000 and < 50,000 patrons
>= 50,000 and < 100,000 patrons
>= 100,000 and < 1,000,000 patrons
>= 1,000,000 patrons
11
24
11
19 5
4
1
1
$ 1,
1,
2,
6,
2,
31,
17,
1, 159
521
.523
882
150
500
500
750
265
$953
2
2 9
5
4
3
3 1
$
500
8,875
47,833
13,000
37,813
24,167
6,000
300,000
$
1,021
-6,352
-40,952
-10,850
-6,313
.088
.597
.251
.431
.825
-$46,881
.193
Not
e,
n = number of responses; ave = mean cost; d = difference in cost between sample groups; p
probability based on Welch's t-test for samples with unequal variances.
two-tailed
Appendix C: Glossary of Software Classified by Category of Intellectual Property
Free software is software that may be viewed, used, copied, modified, and
redistributed. The source code is open to view. These freedoms are ensured by a
reciprocal or "copyleft" license, which extends the freedoms to copies, identical or
modified, of the software. Free software cannot be included in proprietary software.
Any software that includes free software must also be free software. There is no
restriction on what may be charged or not charged for free software; it may be and often
is distributed at no cost. GNU/Linux is an example of free software.
Freeware, not to be confused with free software, is software that costs nothing.
It is distributed in binary form without the source code, often available by download
from the Web. Freeware may be proprietary software, copyrighted and licensed. Users
are restricted from modifying the software, although redistribution may be permitted.
The Adobe Acrobat Reader is an example of freeware.
Open source software (OSS) is for practical purposes the same as free software.
Open source and free software are often licensed by the same licenses, such as the GNU
General Public License. However, some open source licenses do not ensure all the
freedoms that free software licenses do, such as the reciprocal clause. The difference
between open source and free software is mostly one of emphasis. Free software
emphasizes the user's right to freedom, and open source software emphasizes viewable
source code. Sun Microsystem's Java is an example of open source software.
Patented software is software that contains a method or algorithm, the exclusive
right to which is granted to the inventor by the state for a period of time (in the United
93
States, 20 years). A patent protects a method or algorithm in the software, whereas a
copyright protects an expression of an algorithm. The criteria on which a patent is
granted hold the invention to a higher standard than a copyright: utility, novelty, and
nonobviousness. A patent excludes others from making, using, or selling instances of
the invention. An inventor may sell licenses of these rights to manufacturers who may
in turn pay royalty fees to the inventor for each item sold. Examples of software patents
include media compression algorithms such as MP3 audio and DVD video compression.
A patent would not generally apply to an ILS, because an ILS is commodified software,
that is, the functions of all ILSs are fairly homogeneous and well known.
Proprietary software is owned by someone who has the exclusive right to
distribute the software. The software is copyrighted. Transfer of use of the software is
by a license, often called an End User License Agreement (EULA), which is a legal
contract that specifies the terms of use. A license to use proprietary software generally
does not permit copying the software. The software is distributed in compiled binary
form and the source code is not revealed to users. Microsoft Word is an example of
proprietary software.
Public domain software is the simplest type of software as intellectual property.
This software is not copyrighted and has no restrictions. Copies or modified copies of
public domain software can be copyrighted. Thus public domain software can be taken
out of the public domain and made proprietary. By the Berne Convention (World
Intellectual Property Organization, 1979), which is an international treaty regarding
copyright, exclusive rights are automatically granted to authors as soon as a work is
94
fixed (e.g., written or recorded) in a physical medium. Therefore, to place a work in the
public domain requires that the author forfeits rights to it or that the copyright expires.
For example, Megginson (2000) wrote a group of subroutines for XML and placed it in
the public domain by disclaiming his rights to it (Apodaca, 2006). Sun Microsystems
(2004) incorporated it into their Java product with no obligation to Megginson.
Shareware is the same as freeware, except that shareware is available for a
limited time period at no cost, after which it ceases to function unless a payment is made.
WinZip 11.0 file compression software, which has a 45-day evaluation period, is an
example of shareware.
95
top related